| { |
| "paper_id": "2021", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T01:10:51.840254Z" |
| }, |
| "title": "Exploratory analysis of news sentiment using subgroup discovery", |
| "authors": [ |
| { |
| "first": "Anita", |
| "middle": [], |
| "last": "Valmarska", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "University of Ljubljana", |
| "location": { |
| "postCode": "113", |
| "settlement": "Ljubljana", |
| "country": "Slovenia" |
| } |
| }, |
| "email": "anita.valmarska@ijs.si" |
| }, |
| { |
| "first": "Luis", |
| "middle": [], |
| "last": "Adri\u00e1n", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "L3i laboratory University", |
| "institution": "", |
| "location": { |
| "addrLine": "of La Rochelle La Rochelle", |
| "settlement": "luis.cabrera diego, elvys.linhares pontes", |
| "country": "France" |
| } |
| }, |
| "email": "" |
| }, |
| { |
| "first": "Elvys", |
| "middle": [ |
| "Linhares" |
| ], |
| "last": "Pontes", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "L3i laboratory University", |
| "institution": "", |
| "location": { |
| "addrLine": "of La Rochelle La Rochelle", |
| "settlement": "luis.cabrera diego, elvys.linhares pontes", |
| "country": "France" |
| } |
| }, |
| "email": "" |
| }, |
| { |
| "first": "Senja", |
| "middle": [], |
| "last": "Pollak", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "L3i laboratory University", |
| "institution": "", |
| "location": { |
| "addrLine": "of La Rochelle La Rochelle", |
| "settlement": "luis.cabrera diego, elvys.linhares pontes", |
| "country": "France" |
| } |
| }, |
| "email": "senja.pollak@ijs.si" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "In this study, we present an exploratory analysis of a Slovenian news corpus, in which we investigate the association between named entities and sentiment in the news. We propose a methodology that combines Named Entity Recognition and Subgroup Discovery-a descriptive rule learning technique for identifying groups of examples that share the same class label (sentiment) and pattern (features-Named Entities). The approach is used to induce the positive and negative sentiment class rules that reveal interesting patterns related to different Slovenian and international politicians, organizations, and locations.", |
| "pdf_parse": { |
| "paper_id": "2021", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "In this study, we present an exploratory analysis of a Slovenian news corpus, in which we investigate the association between named entities and sentiment in the news. We propose a methodology that combines Named Entity Recognition and Subgroup Discovery-a descriptive rule learning technique for identifying groups of examples that share the same class label (sentiment) and pattern (features-Named Entities). The approach is used to induce the positive and negative sentiment class rules that reveal interesting patterns related to different Slovenian and international politicians, organizations, and locations.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "Traditionally, sentiment analysis refers to the use of natural language processing to systematically identify, extract, quantify, and study affective states and subjective information. Most frequently, it is used as a predictive technique used to model social media (Beigi et al., 2016) , more specifically to predict or summarize opinions, attitudes and emotions in tweets, comments, online reviews etc., where the main focus is on predicting attitudes expressed towards a specific entity (Mejova, 2009) . Another line of research applies sentiment analysis on news text, where the focus has shifted from analyzing sentiment towards a specific target to analyzing the intrinsic mood of the text itself (Pelicon et al., 2020) . Authors aimed to model feelings (positive, negative, or neutral) that readers feel while reading a certain piece of news (Bu\u010dar et al., 2018; Liu, 2012; Pelicon et al., 2020) , also in relation to news covering Covid-19 (Aslam et al., 2020) , modelled news sentiment in relation to stock market and economic conditions (Van de Kauter et al., 2015; Bowden et al., 2019; Rambaccussing and Kwiatkowski, 2020) . Sentiment analysis has been also used in fake news identification (Bhutani et al., 2019) and in media bias analysis (El Ali et al., 2018) .", |
| "cite_spans": [ |
| { |
| "start": 266, |
| "end": 286, |
| "text": "(Beigi et al., 2016)", |
| "ref_id": "BIBREF1" |
| }, |
| { |
| "start": 490, |
| "end": 504, |
| "text": "(Mejova, 2009)", |
| "ref_id": "BIBREF18" |
| }, |
| { |
| "start": 703, |
| "end": 725, |
| "text": "(Pelicon et al., 2020)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 849, |
| "end": 869, |
| "text": "(Bu\u010dar et al., 2018;", |
| "ref_id": "BIBREF6" |
| }, |
| { |
| "start": 870, |
| "end": 880, |
| "text": "Liu, 2012;", |
| "ref_id": "BIBREF16" |
| }, |
| { |
| "start": 881, |
| "end": 902, |
| "text": "Pelicon et al., 2020)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 948, |
| "end": 968, |
| "text": "(Aslam et al., 2020)", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 1047, |
| "end": 1075, |
| "text": "(Van de Kauter et al., 2015;", |
| "ref_id": "BIBREF11" |
| }, |
| { |
| "start": 1076, |
| "end": 1096, |
| "text": "Bowden et al., 2019;", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 1097, |
| "end": 1133, |
| "text": "Rambaccussing and Kwiatkowski, 2020)", |
| "ref_id": "BIBREF22" |
| }, |
| { |
| "start": 1202, |
| "end": 1224, |
| "text": "(Bhutani et al., 2019)", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 1252, |
| "end": 1273, |
| "text": "(El Ali et al., 2018)", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "In the current trend of natural language processing research (Rogers and Augenstein, 2020) , the main focus is on improving the predictive performance over state-of-the art especially using deep learning-based methods. The drawback of these models is in their very limited interpretability. In contrast, several data and text mining techniques have been developed to improve domain understanding and support exploratory analysis of data, with focus on explainable models, which is crucial e.g. in medical applications, but also interesting for interdisciplinary research in the field of digital humanities and digital social sciences. Our research falls under this line of research.", |
| "cite_spans": [ |
| { |
| "start": 61, |
| "end": 90, |
| "text": "(Rogers and Augenstein, 2020)", |
| "ref_id": "BIBREF23" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "The aim of our study is to gain better understanding into news sentiment by analysis of named entities in a manually annotated corpus of Slovenian news articles (Bu\u010dar, 2017) . More specifically, our aim is to identify groups of topics with negative or positive sentiment in Slovenian news, where topics are identified by named entities and their interaction forms the context of the reported stories. We propose the employment of subgroup discovery -a descriptive rule learning technique for identification of groups of examples sharing the same class label (sentiment) and same pattern (features). The task of subgroup discovery is the combination of predictive and descriptive rule induction. The result of subgroup discovery is to provide understandable descriptions of subgroups of individuals which share a common target property of interest. Subgroup discovery methods have traditionally be successfully applied to in different medical applications (e.g. detecting of groups of patients at risk for atherosclerotic cardiovascular disease (Gamberger and Lavra\u010d, 2002) , supporting factors for brain ischemia (Gamberger and Lavra\u010d, 2007) , and psychiatric emergency (Carmona et al., 2011 ), but only rarely applied to model textual data. The closest to our study is the work by (Vavpeti\u010d et al., 2013) using the subgroup discovery system Hedwig for analyzing news articles about Portugal focusing on interesting vocabulary patterns that reflect credit default swap. The authors focused on financial entities, geographical entities and a specialized vocabulary of the European sovereign debt crisis.", |
| "cite_spans": [ |
| { |
| "start": 161, |
| "end": 174, |
| "text": "(Bu\u010dar, 2017)", |
| "ref_id": "BIBREF5" |
| }, |
| { |
| "start": 1045, |
| "end": 1073, |
| "text": "(Gamberger and Lavra\u010d, 2002)", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 1114, |
| "end": 1142, |
| "text": "(Gamberger and Lavra\u010d, 2007)", |
| "ref_id": "BIBREF9" |
| }, |
| { |
| "start": 1171, |
| "end": 1192, |
| "text": "(Carmona et al., 2011", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 1283, |
| "end": 1306, |
| "text": "(Vavpeti\u010d et al., 2013)", |
| "ref_id": "BIBREF27" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "The main contributions of this paper are two fold. First, we propose a novel approach using named entity recognition and linking in a subgroup discovery setting. Next, we apply the method on the Slovenian news dataset, getting new insights into Slovenian news reporting in terms of news sentiment and showcase the potential of our approach for digital social science research.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "The paper is structured as follows: in Section 2, we present the data used in the experimental work. Section 3 presents a short outline of the employed methodology. In Section 4 and Section 5, we present our results and offer our conclusions and ideas for further work.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "In our experiments, we used the manually sentiment annotated Slovenian news corpus SentiNews 1.0 (Bu\u010dar et al., 2018; Bu\u010dar, 2017) 1 . The corpus consists of Slovene web-crawled news containing more than 250,000 documents with political, business, economic and financial content from five Slovene media resources on the web. The data covers the period between 1 September 2007 to 31 December 2013. Data used in the experiments is a manually sentiment annotated stratified random sample of 10,427 documents from news portals 24ur, Dnevnik, Finance, Rtvslo, and\u017durnal24. Data was independently annotated by 2-6 annotators, using the five-level Lickert scale (1 -very negative, 2 -negative, 3 -neutral, 4 -positive, and 5 -very positive) on three levels of granularity, i.e. on document, paragraph, and sentence level. The sentiment of an instance is defined as the average of the sentiment scores given by the different annotators, where an instance labeled as negative has received an average score less than or equal to 2.4 and an instance labeled as positive has received an average score to 3.6. Instances with an average score in-between were labeled as neutral.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data", |
| "sec_num": "2" |
| }, |
| { |
| "text": "The analysis of the agreement between annotators is available in (Bu\u010dar et al., 2018) . The value of annotators agreement on document level as measured by the Cronbach's alpha is 0.903.", |
| "cite_spans": [ |
| { |
| "start": 65, |
| "end": 85, |
| "text": "(Bu\u010dar et al., 2018)", |
| "ref_id": "BIBREF6" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data", |
| "sec_num": "2" |
| }, |
| { |
| "text": "In this paper we are interested only in documents with either positive or negative sentiment, which corresponds to 1665 positive and 3337 negative articles, respectively. Note that the dataset is thus imbalanced towards the negative class (which is also matching the observations of media researchers that attention to negative news is disproportionate (e.g. (Van der Meer et al., 2019; Soroka et al., 2019) ).", |
| "cite_spans": [ |
| { |
| "start": 359, |
| "end": 386, |
| "text": "(Van der Meer et al., 2019;", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 387, |
| "end": 407, |
| "text": "Soroka et al., 2019)", |
| "ref_id": "BIBREF24" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data", |
| "sec_num": "2" |
| }, |
| { |
| "text": "The methodology for named entity-based sentiment subgroup discovery consists of three steps.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Methodology", |
| "sec_num": "3" |
| }, |
| { |
| "text": "For each document from the corpora, we perform named entity recognition (NER) and named entity linking (NEL) using the approaches described in Boros et al. (2020) 2and Linhares Pontes et al. (2020) 3 , respectively. Specifically, for the NER system we fine-tuned CroSloEngual BERT (Ul\u010dar and Robnik-\u0160ikonja, 2020) with two staked Transformer blocks on the top. For the NEL system we used the architecture founded on the Multilingual End-to-End Entity Linking with match correction and candidate filtering. Both systems, NER and NEL, were trained using the Slovene WikiANN dataset (Pan et al., 2017) . The dataset was split in three partitions, train, development and test. The evaluation on the test partition, showed that the NER system has a micro F-score of 0.954, while the NEL system has an F-score of 0.705.", |
| "cite_spans": [ |
| { |
| "start": 580, |
| "end": 598, |
| "text": "(Pan et al., 2017)", |
| "ref_id": "BIBREF19" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Named entity recognition and linking", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "In SentiNews 1.0, we identified 914 person names, 699 organizations and 476 locations with assigned NEL identifiers. We used the NEL codes to extract the nominative case of the named entities from the the Slovenian Wikipedia.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Named entity recognition and linking", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "As the state-of-the-art algorithms for subgroup discovery work on structured data, the second step of the methodology is to transform the discovered (and linked) named entities from step 1 into a tabular form suitable for subgroup discovery. The resulting tables were constructed by representing 2 Code available at:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data transformation", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "https://github.com/ EMBEDDIA/stacked-ner 3 Code available at: https://github.com/ EMBEDDIA/multilingual_entity_linking each non-neutral sentiment document from the corpora as a row in a table. The documents are described by the values of the identified entities, yes if the respective entity was identified in the document and no if the entity was not present in the document. The document's sentiment represent the class label.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data transformation", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "The result of data transformation is a table with 2645 rows (i.e. documents with positive or negative class label, and identified linked named entity) and 2089 columns (i.e. named entities corresponding to person, organisation and location names) as attributes. In our order to reduce the number of columns and improve the chances of the subgroup discovery algorithm of discovering good subgroups, we chose to proceed only with top n (with n=20) most frequent entities from each entity group: person, organization, and location. 4 As we are only interested into entities mentioned in the articles, we removed all documents without any identified entity, thus resulting in a table with 1703 rows and 60 columns i.e. entities, with the positive and negative sentiment class distribution (560, 1143), which is similar to the dataset distribution before preprocessing). Table 1 presents the 20 most frequent persons, organizations, and locations identified in the corpora. Entities are ordered according to their frequency, the most popular entities are written on top.", |
| "cite_spans": [ |
| { |
| "start": 529, |
| "end": 530, |
| "text": "4", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 866, |
| "end": 873, |
| "text": "Table 1", |
| "ref_id": "TABREF1" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Data transformation", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "The third step of the methodology is the identification of subgroups via subgroup discovery. Subgroup discovery is a descriptive induction technique that learns descriptive rules from labeled data. The task of subgroup discovery is to find interesting subgroups in the population, i.e. subgroups that have a significantly different class distribution than the entire population (Kl\u00f6sgen, 1996; Wrobel, 1997) . The result of subgroup discovery is a set of individual rules, where the rule consequence is a class label. An important characteristic of subgroup discovery is that its task is a combination of predictive and descriptive rule induction. It provides understandable descriptions of subgroups of individuals which share a common target property of interest.", |
| "cite_spans": [ |
| { |
| "start": 378, |
| "end": 393, |
| "text": "(Kl\u00f6sgen, 1996;", |
| "ref_id": "BIBREF12" |
| }, |
| { |
| "start": 394, |
| "end": 407, |
| "text": "Wrobel, 1997)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Subgroup discovery", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "We performed subgroup discovery using the DoubleBeam-SD algorithm (Valmarska et 2017) 5 . The DoubleBeam-SD subgroup discovery algorithm combines separate refinement and selection heuristics with the beam search. Inverted m-estimate is used as the heuristics in the rule refinement phase while m-estimate is the rule selection heuristics. The width of the beam was set to 100. The algorithm was set to extract up to 10 best rules for each class. Table 2 presents the rules describing groups of documents with positive or negative sentiment. The prefixes PER , ORG , and LOC denote a person, an organization or a location respectively. For example, the interpretation of the first rule, LOC Nova Gorica = no AND PER Igor Bav\u010dar = yes AND LOC Grad Brdo pri Kranju = no \u2192 negative is as follows: articles that talk about the person Igor Bav\u010dar and do not talk about the locations Nova Gorica and Grad Brdo pri Kranju (corresponding to Brdo pri Kranju, the location of Slovenian government's main venue for diplomatic meetings) are articles with negative sentiment. This rule covers 28 documents, out of which 26 are actually labeled with negative sentiment, while 2 are annotated with positive sentiment.", |
| "cite_spans": [ |
| { |
| "start": 66, |
| "end": 79, |
| "text": "(Valmarska et", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 446, |
| "end": 453, |
| "text": "Table 2", |
| "ref_id": "TABREF3" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Subgroup discovery", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "Rules describing articles with negative sentiment concern articles written about Igor Bav\u010dar, a Slovenian politician and manager. As politician, he played an important role during the Slovenian independence period (he acted as minister of internal affairs in the beginning of the 90s and later as Slovenian minister for European affairs) (Plut-Pregelj et al., 2018) . In 2002, he withdraw from politics to engage in the private sector and was connected to financial affairs, which explains the negative sentiment association in our corpus. Bav\u010dar became the chairman of the Istrabenz holding company in 2002. On 31 March 2009 Bav\u010dar resigned as the President of Istrabenz due to poor business performance as the company was forced into financial restructuring. Bav\u010dar was indicted for disputed trading in Istrabenz shares in 2017, when he attempted to take over the company. He was arrested in financial fraud investigation and released after 10 hours. 6 The discovered subgroups of articles with positive sentiment contain articles talking either about Slovenia's former president, Danilo T\u00fcrk, or the city of New York, New Yorku. Danilo T\u00fcrk is a Slovenian diplomat, professor of international law, human rights expert, and political figure who served as President of Slovenia from 2007 to 2012. Note that the rules of discovered subgroups have a low precision (as the distribution of true positive and false positive is 53% vs. 47%). However one should take into account that the dataset is very imbalanced -with only 33% of articles annotated with a positive sentiment class label. These rules also show the main objective of subgroup discovery -discovering and describing groups of examples with significantly different class distribution to the the population class distribution rather than prediction.", |
| "cite_spans": [ |
| { |
| "start": 338, |
| "end": 365, |
| "text": "(Plut-Pregelj et al., 2018)", |
| "ref_id": "BIBREF21" |
| }, |
| { |
| "start": 953, |
| "end": 954, |
| "text": "6", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "4" |
| }, |
| { |
| "text": "It is interesting that the positive sentiment for New York is connected to the absence of other named entities, closely related to New York including the former president of the United States, Barack Obama. Another entity whose omission in news articles strengthens the positive senti-6 https://en.wikipedia.org/wiki/Igor_ Bav%C4%8Dar, Last accessed: 8 February 2021. ment associated with New York is the investment bank Lehman Brothers. On September 15, 2008, Lehman Brothers Holdings, Inc. sought Chapter 11 protection, initiating the largest bankruptcy proceeding in United States (U.S.) history. It declared $639 billion in assets and $613 billion in debts. At the time, Lehman was the fourth-largest U.S. investment bank. Despite being thought \"too big to fail\", the federal government did not employ extraordinary measures to save Lehman, such as the enabling financing it had facilitated for J.P. Morgan Chase's purchase of a failing Bear Stearns just six months earlier. Lehman's demise was a seminal event in the financial crisis that began in the U.S. subprime mortgage industry in 2007, spread to the credit markets, and then burned through the world's financial markets. The crisis resulted in significant and wide losses to the economy. (Wiggins et al., 2014) Other rules of positive sentiment with named entity New York, also include the absence of Standard & Poor's, an American credit rating agency and a division of S&P Global (ORG Standard & Pools) that publishes financial research and analysis on stocks, bonds, and commodities, and the Bank of America (ORG Bank of America).", |
| "cite_spans": [ |
| { |
| "start": 1250, |
| "end": 1272, |
| "text": "(Wiggins et al., 2014)", |
| "ref_id": "BIBREF28" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The positive sentiment of Danilo T\u00fcrk is related to the omission of the above mentioned politician involved in financial affairs, Igor Bav\u010dar. The quality of the rules presented in Table 2 is presented in Table 3 in the appendix.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 181, |
| "end": 188, |
| "text": "Table 2", |
| "ref_id": "TABREF3" |
| }, |
| { |
| "start": 205, |
| "end": 212, |
| "text": "Table 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "4" |
| }, |
| { |
| "text": "In this paper, we presented a sentiment subgroup discovery approach through the lenses of named entities. Experiments were performed on a corpora of Slovene news data, manually labeled with positive and negative sentiment. Our approach utilizes the fact that the sentiment in news articles is tied-up to their discussed named entities. We identified groups of news articles with negative and positive sentiment.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusions", |
| "sec_num": "5" |
| }, |
| { |
| "text": "The identified groups of articles were described by rules containing only one positive condition, confirming the presence of a named entity in an article. Many of the identified subgroups for a chosen sentiment focus on the same entity. One of the key person entities associated with negative sentiment is Igor Bav\u010dar -a polarizing Slovenian politician and manager, whose presence in the news from 2009 is accompanied by negative sentiment due to financial affairs. We are also able to identify a generally positive sentiment towards another Slovenian politician and a location.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusions", |
| "sec_num": "5" |
| }, |
| { |
| "text": "In future work we will adapt the used algorithm for subgroup discovery to identify subgroups that are not overlapping, thus potentially involving multiple entities. A Appendix. Quality of subgroup rules Table 3 presents the quality measures of rules describing subgroups presented in Table 2 . Note that the purpose of these experiments is exploratory analysis, not a prediction classification task and that the results are reported on the entire dataset, without any train-test data split. In addition to the traditional measures of precision, accuracy and recall, we also present values of weighted relative accuracy (WRACC). Table 3 : Quality of the rules describing subgroup presented in Table 2 . Lavra\u010d et al. (2004) argue that WRACC, also referred to as unusualness, is the most important measure of subgroup discovery rules. WRACC for a chosen rule is defined as follows W RACC = T P + F P P + N { T P T P + F P \u2212 P P + N },", |
| "cite_spans": [ |
| { |
| "start": 702, |
| "end": 722, |
| "text": "Lavra\u010d et al. (2004)", |
| "ref_id": "BIBREF13" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 203, |
| "end": 210, |
| "text": "Table 3", |
| "ref_id": null |
| }, |
| { |
| "start": 284, |
| "end": 291, |
| "text": "Table 2", |
| "ref_id": "TABREF3" |
| }, |
| { |
| "start": 628, |
| "end": 635, |
| "text": "Table 3", |
| "ref_id": null |
| }, |
| { |
| "start": 692, |
| "end": 699, |
| "text": "Table 2", |
| "ref_id": "TABREF3" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Conclusions", |
| "sec_num": "5" |
| }, |
| { |
| "text": "(1) where TP is the number of true positive examples covered by the rule, FP is the number of false positive examples covered by the rule, while P and N are the number of positive and negative examples, respectively, in the population. WRACC reflects both the rule significance and rule coverage, as subgroup discovery is interested in rules with significantly different class distribution than the prior class distribution that covers many instances.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusions", |
| "sec_num": "5" |
| }, |
| { |
| "text": "Data is available on https://www.clarin.si/ repository/xmlui/handle/11356/1110", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "The number of most frequent named entities (n) from each category was set arbitrarily.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "The code of the DoubleBeam-SD algorithm is available on https://github.com/bib3rce/RL_SD.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "This work was supported by the Slovenian Research Agency (ARRS) grants for the programmes, Knowledge technologies (P2-0103) and Artificial intelligence and inteligent systems (P2-0209), the project Computer-assisted multilingual news discourse analysis with contextual embeddings (CAN-DAS, J6-2581), as well as the European Union's Horizon 2020 research and innovation programme under grant agreement No 825153, project EM-BEDDIA (Cross-Lingual Embeddings for Less-Represented Languages in European News Media).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acknowledgments", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Sentiments and emotions evoked by news headlines of coronavirus disease (covid-19) outbreak", |
| "authors": [ |
| { |
| "first": "Faheem", |
| "middle": [], |
| "last": "Aslam", |
| "suffix": "" |
| }, |
| { |
| "first": "Jabir", |
| "middle": [], |
| "last": "Tahir Mumtaz Awan", |
| "suffix": "" |
| }, |
| { |
| "first": "Aisha", |
| "middle": [], |
| "last": "Hussain Syed", |
| "suffix": "" |
| }, |
| { |
| "first": "Mahwish", |
| "middle": [], |
| "last": "Kashif", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Parveen", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Humanities and Social Sciences Communications", |
| "volume": "7", |
| "issue": "1", |
| "pages": "1--9", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Faheem Aslam, Tahir Mumtaz Awan, Jabir Hussain Syed, Aisha Kashif, and Mahwish Parveen. 2020. Sentiments and emotions evoked by news headlines of coronavirus disease (covid-19) outbreak. Human- ities and Social Sciences Communications, 7(1):1-9.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "An overview of sentiment analysis in social media and its applications in disaster relief", |
| "authors": [ |
| { |
| "first": "Ghazaleh", |
| "middle": [], |
| "last": "Beigi", |
| "suffix": "" |
| }, |
| { |
| "first": "Xia", |
| "middle": [], |
| "last": "Hu", |
| "suffix": "" |
| }, |
| { |
| "first": "Ross", |
| "middle": [], |
| "last": "Maciejewski", |
| "suffix": "" |
| }, |
| { |
| "first": "Huan", |
| "middle": [], |
| "last": "Liu", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Sentiment analysis and ontology engineering", |
| "volume": "", |
| "issue": "", |
| "pages": "313--340", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ghazaleh Beigi, Xia Hu, Ross Maciejewski, and Huan Liu. 2016. An overview of sentiment analysis in social media and its applications in disaster relief. Sentiment analysis and ontology engineering, pages 313-340.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Fake news detection using sentiment analysis", |
| "authors": [ |
| { |
| "first": "Bhavika", |
| "middle": [], |
| "last": "Bhutani", |
| "suffix": "" |
| }, |
| { |
| "first": "Neha", |
| "middle": [], |
| "last": "Rastogi", |
| "suffix": "" |
| }, |
| { |
| "first": "Priyanshu", |
| "middle": [], |
| "last": "Sehgal", |
| "suffix": "" |
| }, |
| { |
| "first": "Archana", |
| "middle": [], |
| "last": "Purwar", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Twelfth International Conference on Contemporary Computing (IC3)", |
| "volume": "", |
| "issue": "", |
| "pages": "1--5", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bhavika Bhutani, Neha Rastogi, Priyanshu Sehgal, and Archana Purwar. 2019. Fake news detection using sentiment analysis. In 2019 Twelfth Inter- national Conference on Contemporary Computing (IC3), pages 1-5. IEEE.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Alleviating Digitization Errors in Named Entity Recognition for Historical Documents", |
| "authors": [ |
| { |
| "first": "Emanuela", |
| "middle": [], |
| "last": "Boros", |
| "suffix": "" |
| }, |
| { |
| "first": "Ahmed", |
| "middle": [], |
| "last": "Hamdi", |
| "suffix": "" |
| }, |
| { |
| "first": "Elvys", |
| "middle": [ |
| "Linhares" |
| ], |
| "last": "Pontes", |
| "suffix": "" |
| }, |
| { |
| "first": "Luis", |
| "middle": [ |
| "Adri\u00e1n" |
| ], |
| "last": "Cabrera-Diego", |
| "suffix": "" |
| }, |
| { |
| "first": "Jose", |
| "middle": [ |
| "G" |
| ], |
| "last": "Moreno", |
| "suffix": "" |
| }, |
| { |
| "first": "Nicolas", |
| "middle": [], |
| "last": "Sidere", |
| "suffix": "" |
| }, |
| { |
| "first": "Antoine", |
| "middle": [], |
| "last": "Doucet", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Proceedings of the 24th Conference on Computational Natural Language Learning (CoNLL)", |
| "volume": "", |
| "issue": "", |
| "pages": "431--441", |
| "other_ids": { |
| "DOI": [ |
| "10.18653/v1/2020.conll-1.35" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Emanuela Boros, Ahmed Hamdi, Elvys Lin- hares Pontes, Luis Adri\u00e1n Cabrera-Diego, Jose G. Moreno, Nicolas Sidere, and Antoine Doucet. 2020. Alleviating Digitization Errors in Named Entity Recognition for Historical Documents. In Proceedings of the 24th Conference on Computa- tional Natural Language Learning (CoNLL), pages 431-441, Online. Association for Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Economy through a lens: Distortions of policy coverage in uk national newspapers", |
| "authors": [ |
| { |
| "first": "James", |
| "middle": [], |
| "last": "Bowden", |
| "suffix": "" |
| }, |
| { |
| "first": "Andrzej", |
| "middle": [], |
| "last": "Kwiatkowski", |
| "suffix": "" |
| }, |
| { |
| "first": "Dooruj", |
| "middle": [], |
| "last": "Rambaccussing", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Journal of Comparative Economics", |
| "volume": "47", |
| "issue": "4", |
| "pages": "881--906", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "James Bowden, Andrzej Kwiatkowski, and Dooruj Rambaccussing. 2019. Economy through a lens: Distortions of policy coverage in uk national newspapers. Journal of Comparative Economics, 47(4):881-906.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Manually sentiment annotated slovenian news corpus SentiNews 1.0. Slovenian language resource repository CLARIN", |
| "authors": [ |
| { |
| "first": "Jo\u017ee", |
| "middle": [], |
| "last": "Bu\u010dar", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jo\u017ee Bu\u010dar. 2017. Manually sentiment annotated slove- nian news corpus SentiNews 1.0. Slovenian lan- guage resource repository CLARIN.SI.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Annotated news corpora and a lexicon for sentiment analysis in slovene. Language Resources and Evaluation", |
| "authors": [ |
| { |
| "first": "Jo\u017ee", |
| "middle": [], |
| "last": "Bu\u010dar", |
| "suffix": "" |
| }, |
| { |
| "first": "Janez", |
| "middle": [], |
| "last": "Martin\u017enidar\u0161i\u010d", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Povh", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "", |
| "volume": "52", |
| "issue": "", |
| "pages": "895--919", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jo\u017ee Bu\u010dar, Martin\u017dnidar\u0161i\u010d, and Janez Povh. 2018. Annotated news corpora and a lexicon for sentiment analysis in slovene. Language Resources and Eval- uation, 52(3):895-919.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Evolutionary fuzzy rule extraction for subgroup discovery in a psychiatric emergency department", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Crist\u00f3bal", |
| "suffix": "" |
| }, |
| { |
| "first": "Pedro", |
| "middle": [], |
| "last": "Carmona", |
| "suffix": "" |
| }, |
| { |
| "first": "Mar\u00eda Jos\u00e9 Del", |
| "middle": [], |
| "last": "Gonz\u00e1lez", |
| "suffix": "" |
| }, |
| { |
| "first": "Mercedes", |
| "middle": [], |
| "last": "Jesus", |
| "suffix": "" |
| }, |
| { |
| "first": "Luis", |
| "middle": [], |
| "last": "Nav\u00edo-Acosta", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Jim\u00e9nez-Trevino", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Soft Computing", |
| "volume": "15", |
| "issue": "12", |
| "pages": "2435--2448", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Crist\u00f3bal J Carmona, Pedro Gonz\u00e1lez, Mar\u00eda Jos\u00e9 del Jesus, Mercedes Nav\u00edo-Acosta, and Luis Jim\u00e9nez- Trevino. 2011. Evolutionary fuzzy rule extraction for subgroup discovery in a psychiatric emergency department. Soft Computing, 15(12):2435-2448.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Measuring, understanding, and classifying news media sympathy on twitter after crisis events", |
| "authors": [ |
| { |
| "first": "Abdallah", |
| "middle": [ |
| "El" |
| ], |
| "last": "Ali", |
| "suffix": "" |
| }, |
| { |
| "first": "Tim", |
| "middle": [ |
| "C" |
| ], |
| "last": "Stratmann", |
| "suffix": "" |
| }, |
| { |
| "first": "Souneil", |
| "middle": [], |
| "last": "Park", |
| "suffix": "" |
| }, |
| { |
| "first": "Johannes", |
| "middle": [], |
| "last": "Sch\u00f6ning", |
| "suffix": "" |
| }, |
| { |
| "first": "Wilko", |
| "middle": [], |
| "last": "Heuten", |
| "suffix": "" |
| }, |
| { |
| "first": "Susanne Cj", |
| "middle": [], |
| "last": "Boll", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems", |
| "volume": "", |
| "issue": "", |
| "pages": "1--13", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Abdallah El Ali, Tim C Stratmann, Souneil Park, Jo- hannes Sch\u00f6ning, Wilko Heuten, and Susanne CJ Boll. 2018. Measuring, understanding, and classi- fying news media sympathy on twitter after crisis events. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, pages 1- 13.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Supporting factors in descriptive analysis of brain ischaemia", |
| "authors": [ |
| { |
| "first": "Dragan", |
| "middle": [], |
| "last": "Gamberger", |
| "suffix": "" |
| }, |
| { |
| "first": "Nada", |
| "middle": [], |
| "last": "Lavra\u010d", |
| "suffix": "" |
| } |
| ], |
| "year": 2007, |
| "venue": "Conference on Artificial Intelligence in Medicine in Europe", |
| "volume": "", |
| "issue": "", |
| "pages": "155--159", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dragan Gamberger and Nada Lavra\u010d. 2007. Support- ing factors in descriptive analysis of brain ischaemia. In Conference on Artificial Intelligence in Medicine in Europe, pages 155-159. Springer.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "Expertguided subgroup discovery: Methodology and application", |
| "authors": [ |
| { |
| "first": "Dragan", |
| "middle": [], |
| "last": "Gamberger", |
| "suffix": "" |
| }, |
| { |
| "first": "Nada", |
| "middle": [], |
| "last": "Lavra\u010d", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "Journal of Artificial Intelligence Research", |
| "volume": "17", |
| "issue": "", |
| "pages": "501--527", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dragan Gamberger and Nada Lavra\u010d. 2002. Expert- guided subgroup discovery: Methodology and appli- cation. Journal of Artificial Intelligence Research, 17:501-527.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "Fine-grained analysis of explicit and implicit sentiment in financial news articles", |
| "authors": [ |
| { |
| "first": "Marjan", |
| "middle": [], |
| "last": "Van De Kauter", |
| "suffix": "" |
| }, |
| { |
| "first": "Diane", |
| "middle": [], |
| "last": "Breesch", |
| "suffix": "" |
| }, |
| { |
| "first": "V\u00e9ronique", |
| "middle": [], |
| "last": "Hoste", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Expert Systems with applications", |
| "volume": "42", |
| "issue": "11", |
| "pages": "4999--5010", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Marjan Van de Kauter, Diane Breesch, and V\u00e9ronique Hoste. 2015. Fine-grained analysis of explicit and implicit sentiment in financial news articles. Expert Systems with applications, 42(11):4999-5010.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Explora: A multipattern and multistrategy discovery assistant", |
| "authors": [ |
| { |
| "first": "Willi", |
| "middle": [], |
| "last": "Kl\u00f6sgen", |
| "suffix": "" |
| } |
| ], |
| "year": 1996, |
| "venue": "Advances in Knowledge Discovery and Data Mining", |
| "volume": "", |
| "issue": "", |
| "pages": "249--271", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Willi Kl\u00f6sgen. 1996. Explora: A multipattern and mul- tistrategy discovery assistant. In Advances in Knowl- edge Discovery and Data Mining, pages 249-271. AAAI/MIT Press.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "Subgroup discovery with", |
| "authors": [ |
| { |
| "first": "Nada", |
| "middle": [], |
| "last": "Lavra\u010d", |
| "suffix": "" |
| }, |
| { |
| "first": "Branko", |
| "middle": [], |
| "last": "Kav\u0161ek", |
| "suffix": "" |
| }, |
| { |
| "first": "Peter", |
| "middle": [], |
| "last": "Flach", |
| "suffix": "" |
| }, |
| { |
| "first": "Ljupco", |
| "middle": [], |
| "last": "Todorovski", |
| "suffix": "" |
| } |
| ], |
| "year": 2004, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nada Lavra\u010d, Branko Kav\u0161ek, Peter Flach, and Ljupco Todorovski. 2004. Subgroup discovery with cn2-sd.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "Entity Linking for Historical Documents: Challenges and Solutions", |
| "authors": [ |
| { |
| "first": "Luis", |
| "middle": [ |
| "Adri\u00e1n" |
| ], |
| "last": "Elvys Linhares Pontes", |
| "suffix": "" |
| }, |
| { |
| "first": "Jose", |
| "middle": [ |
| "G" |
| ], |
| "last": "Cabrera-Diego", |
| "suffix": "" |
| }, |
| { |
| "first": "Emanuela", |
| "middle": [], |
| "last": "Moreno", |
| "suffix": "" |
| }, |
| { |
| "first": "Ahmed", |
| "middle": [], |
| "last": "Boros", |
| "suffix": "" |
| }, |
| { |
| "first": "Nicolas", |
| "middle": [], |
| "last": "Hamdi", |
| "suffix": "" |
| }, |
| { |
| "first": "Micka\u00ebl", |
| "middle": [], |
| "last": "Sid\u00e8re", |
| "suffix": "" |
| }, |
| { |
| "first": "Antoine", |
| "middle": [], |
| "last": "Coustaty", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Doucet", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Proceedings of the 22nd International Conference on Asia-Pacific Digital Libraries (ICADL 2020)", |
| "volume": "", |
| "issue": "", |
| "pages": "215--231", |
| "other_ids": { |
| "DOI": [ |
| "10.1007/978-3-030-64452-9_19" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Elvys Linhares Pontes, Luis Adri\u00e1n Cabrera-Diego, Jose G. Moreno, Emanuela Boros, Ahmed Hamdi, Nicolas Sid\u00e8re, Micka\u00ebl Coustaty, and Antoine Doucet. 2020. Entity Linking for Historical Docu- ments: Challenges and Solutions. In Proceedings of the 22nd International Conference on Asia-Pacific Digital Libraries (ICADL 2020), pages 215-231, Kyoto, Japan. Springer International Publishing.", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Sentiment analysis and opinion mining. Synthesis lectures on human language technologies", |
| "authors": [ |
| { |
| "first": "Bing", |
| "middle": [], |
| "last": "Liu", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "", |
| "volume": "5", |
| "issue": "", |
| "pages": "1--167", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bing Liu. 2012. Sentiment analysis and opinion min- ing. Synthesis lectures on human language technolo- gies, 5(1):1-167.", |
| "links": null |
| }, |
| "BIBREF17": { |
| "ref_id": "b17", |
| "title": "Mediatization and the disproportionate attention to negative news: The case of airplane crashes", |
| "authors": [ |
| { |
| "first": "Gla", |
| "middle": [], |
| "last": "Toni", |
| "suffix": "" |
| }, |
| { |
| "first": "Anne", |
| "middle": [ |
| "C" |
| ], |
| "last": "Van Der Meer", |
| "suffix": "" |
| }, |
| { |
| "first": "Piet", |
| "middle": [], |
| "last": "Kroon", |
| "suffix": "" |
| }, |
| { |
| "first": "Jeroen", |
| "middle": [], |
| "last": "Verhoeven", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Jonkman", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Journalism Studies", |
| "volume": "20", |
| "issue": "6", |
| "pages": "783--803", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Toni GLA Van der Meer, Anne C Kroon, Piet Ver- hoeven, and Jeroen Jonkman. 2019. Mediatization and the disproportionate attention to negative news: The case of airplane crashes. Journalism Studies, 20(6):783-803.", |
| "links": null |
| }, |
| "BIBREF18": { |
| "ref_id": "b18", |
| "title": "Sentiment analysis: An overview", |
| "authors": [ |
| { |
| "first": "Yelena", |
| "middle": [], |
| "last": "Mejova", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Yelena Mejova. 2009. Sentiment analysis: An overview. University of Iowa, Computer Science De- partment.", |
| "links": null |
| }, |
| "BIBREF19": { |
| "ref_id": "b19", |
| "title": "Crosslingual Name Tagging and Linking for 282 Languages", |
| "authors": [ |
| { |
| "first": "Xiaoman", |
| "middle": [], |
| "last": "Pan", |
| "suffix": "" |
| }, |
| { |
| "first": "Boliang", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| }, |
| { |
| "first": "Jonathan", |
| "middle": [], |
| "last": "May", |
| "suffix": "" |
| }, |
| { |
| "first": "Joel", |
| "middle": [], |
| "last": "Nothman", |
| "suffix": "" |
| }, |
| { |
| "first": "Kevin", |
| "middle": [], |
| "last": "Knight", |
| "suffix": "" |
| }, |
| { |
| "first": "Heng", |
| "middle": [], |
| "last": "Ji", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics", |
| "volume": "1", |
| "issue": "", |
| "pages": "1946--1958", |
| "other_ids": { |
| "DOI": [ |
| "10.18653/v1/P17-1178" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Xiaoman Pan, Boliang Zhang, Jonathan May, Joel Nothman, Kevin Knight, and Heng Ji. 2017. Cross- lingual Name Tagging and Linking for 282 Lan- guages. In Proceedings of the 55th Annual Meet- ing of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1946-1958, Van- couver, Canada. Association for Computational Lin- guistics.", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "Zero-shot learning for cross-lingual news sentiment classification", |
| "authors": [ |
| { |
| "first": "Andra\u017e", |
| "middle": [], |
| "last": "Pelicon", |
| "suffix": "" |
| }, |
| { |
| "first": "Marko", |
| "middle": [], |
| "last": "Pranji\u0107", |
| "suffix": "" |
| }, |
| { |
| "first": "Dragana", |
| "middle": [], |
| "last": "Miljkovi\u0107", |
| "suffix": "" |
| }, |
| { |
| "first": "Senja", |
| "middle": [], |
| "last": "Bla\u017e\u0161krlj", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Pollak", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Applied Sciences", |
| "volume": "10", |
| "issue": "17", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Andra\u017e Pelicon, Marko Pranji\u0107, Dragana Miljkovi\u0107, Bla\u017e\u0160krlj, and Senja Pollak. 2020. Zero-shot learn- ing for cross-lingual news sentiment classification. Applied Sciences, 10(17):5993.", |
| "links": null |
| }, |
| "BIBREF21": { |
| "ref_id": "b21", |
| "title": "Historical Dictionary of Slovenia (Historical Dictionaries of Europe)", |
| "authors": [ |
| { |
| "first": "Leopoldina", |
| "middle": [], |
| "last": "Plut-Pregelj", |
| "suffix": "" |
| }, |
| { |
| "first": "Gregor", |
| "middle": [], |
| "last": "Kranjc", |
| "suffix": "" |
| }, |
| { |
| "first": "\u017darko", |
| "middle": [], |
| "last": "Lazarevi\u0107", |
| "suffix": "" |
| }, |
| { |
| "first": "Carole", |
| "middle": [], |
| "last": "Rogel", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Leopoldina Plut-Pregelj, Gregor Kranjc,\u017darko Lazarevi\u0107, and Carole Rogel. 2018. Historical Dictionary of Slovenia (Historical Dictionaries of Europe). Rowman & Littlefield Publishers.", |
| "links": null |
| }, |
| "BIBREF22": { |
| "ref_id": "b22", |
| "title": "Forecasting with news sentiment: Evidence with uk newspapers", |
| "authors": [ |
| { |
| "first": "Dooruj", |
| "middle": [], |
| "last": "Rambaccussing", |
| "suffix": "" |
| }, |
| { |
| "first": "Andrzej", |
| "middle": [], |
| "last": "Kwiatkowski", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "International Journal of Forecasting", |
| "volume": "36", |
| "issue": "4", |
| "pages": "1501--1516", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dooruj Rambaccussing and Andrzej Kwiatkowski. 2020. Forecasting with news sentiment: Evidence with uk newspapers. International Journal of Fore- casting, 36(4):1501-1516.", |
| "links": null |
| }, |
| "BIBREF23": { |
| "ref_id": "b23", |
| "title": "What can we do to improve peer review in nlp? arXiv preprint", |
| "authors": [ |
| { |
| "first": "Anna", |
| "middle": [], |
| "last": "Rogers", |
| "suffix": "" |
| }, |
| { |
| "first": "Isabelle", |
| "middle": [], |
| "last": "Augenstein", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:2010.03863" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Anna Rogers and Isabelle Augenstein. 2020. What can we do to improve peer review in nlp? arXiv preprint arXiv:2010.03863.", |
| "links": null |
| }, |
| "BIBREF24": { |
| "ref_id": "b24", |
| "title": "Cross-national evidence of a negativity bias in psychophysiological reactions to news", |
| "authors": [ |
| { |
| "first": "Stuart", |
| "middle": [], |
| "last": "Soroka", |
| "suffix": "" |
| }, |
| { |
| "first": "Patrick", |
| "middle": [], |
| "last": "Fournier", |
| "suffix": "" |
| }, |
| { |
| "first": "Lilach", |
| "middle": [], |
| "last": "Nir", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Proceedings of the National Academy of Sciences", |
| "volume": "116", |
| "issue": "", |
| "pages": "18888--18892", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Stuart Soroka, Patrick Fournier, and Lilach Nir. 2019. Cross-national evidence of a negativity bias in psy- chophysiological reactions to news. Proceedings of the National Academy of Sciences, 116(38):18888- 18892.", |
| "links": null |
| }, |
| "BIBREF25": { |
| "ref_id": "b25", |
| "title": "FinEst BERT and CroSloEngual BERT", |
| "authors": [ |
| { |
| "first": "Matej", |
| "middle": [], |
| "last": "Ul\u010dar", |
| "suffix": "" |
| }, |
| { |
| "first": "Marko", |
| "middle": [], |
| "last": "Robnik-\u0160ikonja", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Text, Speech, and Dialogue", |
| "volume": "", |
| "issue": "", |
| "pages": "104--111", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Matej Ul\u010dar and Marko Robnik-\u0160ikonja. 2020. FinEst BERT and CroSloEngual BERT. In Text, Speech, and Dialogue, pages 104-111, Cham. Springer In- ternational Publishing.", |
| "links": null |
| }, |
| "BIBREF26": { |
| "ref_id": "b26", |
| "title": "Refinement and selection heuristics in subgroup discovery and classification rule learning", |
| "authors": [ |
| { |
| "first": "Anita", |
| "middle": [], |
| "last": "Valmarska", |
| "suffix": "" |
| }, |
| { |
| "first": "Nada", |
| "middle": [], |
| "last": "Lavra\u010d", |
| "suffix": "" |
| }, |
| { |
| "first": "Johannes", |
| "middle": [], |
| "last": "F\u00fcrnkranz", |
| "suffix": "" |
| }, |
| { |
| "first": "Marko", |
| "middle": [], |
| "last": "Robnik-\u0160ikonja", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Expert Systems with Applications", |
| "volume": "81", |
| "issue": "", |
| "pages": "147--162", |
| "other_ids": { |
| "DOI": [ |
| "10.1016/j.eswa.2017.03.041" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Anita Valmarska, Nada Lavra\u010d, Johannes F\u00fcrnkranz, and Marko Robnik-\u0160ikonja. 2017. Refinement and selection heuristics in subgroup discovery and clas- sification rule learning. Expert Systems with Appli- cations, 81:147-162.", |
| "links": null |
| }, |
| "BIBREF27": { |
| "ref_id": "b27", |
| "title": "Semantic data mining of financial news articles", |
| "authors": [ |
| { |
| "first": "An\u017ee", |
| "middle": [], |
| "last": "Vavpeti\u010d", |
| "suffix": "" |
| }, |
| { |
| "first": "Petra", |
| "middle": [ |
| "Kralj" |
| ], |
| "last": "Novak", |
| "suffix": "" |
| }, |
| { |
| "first": "Miha", |
| "middle": [], |
| "last": "Gr\u010dar", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "International Conference on Discovery Science", |
| "volume": "", |
| "issue": "", |
| "pages": "294--307", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "An\u017ee Vavpeti\u010d, Petra Kralj Novak, Miha Gr\u010dar, Igor Mozeti\u010d, and Nada Lavra\u010d. 2013. Semantic data mining of financial news articles. In International Conference on Discovery Science, pages 294-307. Springer.", |
| "links": null |
| }, |
| "BIBREF28": { |
| "ref_id": "b28", |
| "title": "The Lehman Brothers bankruptcy a: Overview. Yale program on financial stability case study", |
| "authors": [ |
| { |
| "first": "Rosalind", |
| "middle": [], |
| "last": "Wiggins", |
| "suffix": "" |
| }, |
| { |
| "first": "Thomas", |
| "middle": [], |
| "last": "Piontek", |
| "suffix": "" |
| }, |
| { |
| "first": "Andrew", |
| "middle": [], |
| "last": "Metrick", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rosalind Wiggins, Thomas Piontek, and Andrew Met- rick. 2014. The Lehman Brothers bankruptcy a: Overview. Yale program on financial stability case study.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "TABREF0": { |
| "type_str": "table", |
| "content": "<table><tr><td>Persons</td><td>Organizations</td><td>Locations</td></tr><tr><td>Borut Pahor</td><td>Ljubljanska borza</td><td>Zdru\u017eeno</td></tr><tr><td/><td/><td>kraljestvo</td></tr><tr><td>Janez Jan\u0161a</td><td>Evropska komisija</td><td>Luka Koper</td></tr><tr><td>Dow Jones</td><td>Evropska unija</td><td>New York</td></tr><tr><td>Danilo T\u00fcrk</td><td>Telekom Slovenije</td><td>Nova Gorica</td></tr><tr><td>Igor Bav\u010dar</td><td>Ko\u0161arkarski klub</td><td>Murska Sobota</td></tr><tr><td/><td>Zlatorog</td><td/></tr><tr><td>Karl Erjavec</td><td>Radiotelvizija</td><td>Ljubljanska</td></tr><tr><td/><td>Slovenija</td><td>borza</td></tr><tr><td>Gregor Virant</td><td>Newyor\u0161ka borza</td><td>Slovenj Gradec</td></tr><tr><td>Alenka Bratu\u0161ek</td><td>Luka Koper</td><td>Mestna ob\u010dina</td></tr><tr><td/><td/><td>Ljubljana MOL</td></tr><tr><td>Katarina Kresal</td><td>Adria Airways</td><td>Slovenija</td></tr><tr><td>Nogometni klub</td><td colspan=\"2\">Dow Jones\u010crna gora</td></tr><tr><td>Koper</td><td/><td/></tr><tr><td>Mitja Gaspari</td><td>Nova Ljubljanska</td><td>Novo mesto</td></tr><tr><td/><td>banka NLB</td><td/></tr><tr><td>Angela Merkel</td><td colspan=\"2\">Wall Street\u0160kofja Loka</td></tr><tr><td>Goldman Sachs</td><td>Svetovna banka</td><td>Kranjska Gora</td></tr><tr><td>Gregor Golobi\u010d</td><td>Pop TV</td><td>Bosna in</td></tr><tr><td/><td/><td>Hencegovina</td></tr><tr><td>Nicolas Sarkozy</td><td>General Motors</td><td>Bela krajina</td></tr><tr><td>Ivo Sanader</td><td>Lehman Brothers</td><td>Ju\u017ena Koreja</td></tr><tr><td>Dimitrij Rupel</td><td>Bank of America</td><td>Hrva\u0161ka</td></tr><tr><td>Andrej Bajuk</td><td colspan=\"2\">Republika Slovenija Mestna ob\u010dina</td></tr><tr><td/><td/><td>Maribor MOM</td></tr><tr><td>Janeza Drnov\u0161ka</td><td>Standard & Pool</td><td>Nova Zelandija</td></tr><tr><td>Barack Obama</td><td>Deutsche Bank</td><td>Grad Brdo pri</td></tr><tr><td/><td/><td>Kranju</td></tr></table>", |
| "text": "al.,", |
| "html": null, |
| "num": null |
| }, |
| "TABREF1": { |
| "type_str": "table", |
| "content": "<table/>", |
| "text": "", |
| "html": null, |
| "num": null |
| }, |
| "TABREF3": { |
| "type_str": "table", |
| "content": "<table/>", |
| "text": "List of rules describing subgroups of articles with negative and positive sentiment.", |
| "html": null, |
| "num": null |
| }, |
| "TABREF4": { |
| "type_str": "table", |
| "content": "<table/>", |
| "text": "Stefan Wrobel. 1997. An algorithm for multi-relational discovery of subgroups. In Proceedings of the First European Symposium on Principles of Data Mining and Knowledge Discovery, PKDD 1997, pages 78-87.", |
| "html": null, |
| "num": null |
| } |
| } |
| } |
| } |
