outputs/topic_model_report.md

# Topic Modelling Final Submission Report

Journal: Enterprise Information Systems
Papers analysed: 788
Years: 2007 to 2025

## Method
The model uses one vector per paper from Title + Abstract; DOI retained as paper identifier. Embedding model: allenai/specter2_base. Clustering: UMAP + HDBSCAN (external). The optimizer searches UMAP/HDBSCAN parameters and selects the lowest penalty solution against the required 15-25 clusters, 5 minimum papers per cluster, and 100 maximum papers per cluster.

## Selected Parameters
```json
{
  "algorithm": "UMAP + HDBSCAN (external)",
  "umap_n_neighbors": 10,
  "umap_n_components": 5,
  "umap_metric": "cosine",
  "hdbscan_min_cluster_size": 8,
  "hdbscan_min_samples": null,
  "hdbscan_metric": "euclidean",
  "n_clusters": 23,
  "noise_ratio": 0.0,
  "min_size": 6,
  "max_size": 100,
  "too_small": 0,
  "too_large": 0,
  "silhouette_cosine": 0.082,
  "score": -0.082
}
```

## Validated Clusters
- C21: Process Enterprise Modelling (100 papers, confidence 0.739, PAJAIS: Enterprise Systems and ERP). Evidence titles: Concept-based query language approach to enterprise information systems | Identifying influential factors of business process performance using dependency analysis | Exploiting semantic linkages among multiple sources for semantic information retrieval
- C6: Algorithm Proposed Optimisation (74 papers, confidence 0.797, PAJAIS: Supply Chain and Logistics IS). Evidence titles: Effectiveness of Q-learning as a tool for calibrating agent-based supply network models | Supporting capacity sharing in the cloud manufacturing environment based on game theory and fuzzy logic | Online reinforcement learning-based inventory control for intelligent E-Fulfilment dealing with nonstationary demand
- C17: Erp Enterprise Implementation (54 papers, confidence 0.715, PAJAIS: Enterprise Systems and ERP). Evidence titles: Nursing shortage in the public healthcare system: an exploratory study of Hong Kong | Human factors among workers in a small manufacturing enterprise: a case study | Promoting e-banking actual usage: mix of technology acceptance model and technology-organisation-environment framework
- C4: Social Online Knowledge (52 papers, confidence 0.576, PAJAIS: Social Commerce and Social Media). Evidence titles: Development of an intelligent hospital information chatbot and evaluation of its system usability | Use of neurometrics to choose optimal advertisement method for omnichannel business | Factors influencing the adoption of telemedicine health services during COVID-19 pandemic crisis: an integrative research model
- C15: Manufacturing Cloud Iot (51 papers, confidence 0.62, PAJAIS: Smart Systems IoT and Smart Cities). Evidence titles: Efficient computation of comprehensive statistical information of large OWL datasets: a scalable approach | Transforming cold chain logistics: a reversible vehicle routing approach for sustainable and efficient delivery of perishable goods | FPGA based Industrial Ethernet Network Analyser for Real-time Systems Providing Openness for Industry 4.0
- C14: Supply Supply Chain Chain (50 papers, confidence 0.601, PAJAIS: Supply Chain and Logistics IS). Evidence titles: Internal control system and mechanism of Chinese listed companies: an empirical study | The evolution of creativity: how generative AI is reshaping the hospitality landscape | Credit risk evaluation on technological SMEs in China
- C8: Text Learning Classification (48 papers, confidence 0.586, PAJAIS: Artificial Intelligence and Machine Learning). Evidence titles: Leveraging human-centred AI in interactive English education for generating learning materials | Impact of investor sentiment on mutual fund risk taking and performance: evidence from China | Crowdsourcing-based semantic relation recognition for natural language questions over RDF data
- C16: Enterprise Process Framework (42 papers, confidence 0.501, PAJAIS: Enterprise Systems and ERP). Evidence titles: Application and project portfolio valuation using enterprise architecture and business requirements modelling | New product design and implementation of aboleth: a mobile D&D character creator for enterprise mobile applications and metaverse | A cybersecurity framework for enhancing Small and medium-sized enterprises (SMEs) security posture using user behaviour analytics
- C1: Supply Supply Chain Chain (30 papers, confidence 0.742, PAJAIS: Supply Chain and Logistics IS). Evidence titles: A multi-agent-based model for a negotiation support system in electronic commerce | Renewable energy investment study for electric power enterprise based on a time period with expected supply | Impacts of social networks in an agent-based artificial stock market
- C20: Service Services Web (29 papers, confidence 0.606, PAJAIS: Cloud Computing and SaaS). Evidence titles: Neural adaptive admission control framework: SLA-driven action termination for real-time application service management | Business information query expansion through semantic network | Business process monitoring via decentralized complex event processing
- C0: Blockchain (28 papers, confidence 0.791, PAJAIS: Supply Chain and Logistics IS). Evidence titles: A novel service level agreement model using blockchain and smart contract for cloud manufacturing in industry 4.0 | A framework for rocket and satellite launch information management systems based on blockchain technology | Dynamic trust management framework using blockchain for zero-trust-based authentication in BYOD environments
- C2: Privacy (28 papers, confidence 0.624, PAJAIS: Privacy, Security and Trust). Evidence titles: Piracy in cyber space: Consumer complicity, pirates and enterprise enforcement | Combating the counterfeits with web portal technology | Trust models for efficient communication in Mobile Cloud Computing and their applications to e-Commerce
- C13: Design Method Process (25 papers, confidence 0.616, PAJAIS: Enterprise Systems and ERP). Evidence titles: Identification of approximately duplicate material records in ERP systems | A new attribute reduction method and its application in covering information systems | A systems approach to improve reliability of a contract by Modularising contract’s information flow architecture: a new contribution to risk mitigation in projects management
- C9: Algorithm Learning Intelligent (22 papers, confidence 0.658, PAJAIS: Artificial Intelligence and Machine Learning). Evidence titles: PersonalisedComfort: a personalised thermal comfort model to predict thermal sensation votes for smart building residents | Data driven analysis and forecasting of medium and heavy truck fuel consumption | Modelling omnipresent AI embedding cyber-physical systems by using a novel invariant-based, quantum-inspired fault detection and Bayesian diagnosis approach
- C12: Enterprise Performance Resilience (22 papers, confidence 0.611, PAJAIS: IS Strategy and Management). Evidence titles: Strategic responses to market uncertainty: performance value of strategic agility and online-to-offline platform adoption | AI in airport operations: enhancing competitiveness and satisfaction | SSCM: a secured approach to supply chain management with control management using blowfish optimization
- C3: Proposed Detection Performance (21 papers, confidence 0.766, PAJAIS: Mobile Commerce and Applications). Evidence titles: Adaptive monitoring for autonomous vehicles using the HAFLoop architecture | Design and implementation of embedded un-interruptible power supply system (EUPSS) for web-based mobile application | An intelligent and secure system for predicting and preventing Zika virus outbreak using Fog computing
- C10: Supply Risk Chain (21 papers, confidence 0.481, PAJAIS: Supply Chain and Logistics IS). Evidence titles: Making autonomous vehicle systems human-like: lessons learned from accident experiences in traffic | Analysing and forecasting the security in supply-demand management of Chinese forestry enterprises by linear weighted method and artificial neural network | An effective framework for finding similar cases of dengue from audio and text data using domain thesaurus and case base reasoning
- C19: Modelling Support Domain (19 papers, confidence 0.541, PAJAIS: Enterprise Systems and ERP). Evidence titles: Information visualisation based on graph models | An industrial network flow information integration model for supply chain management and intelligent transportation | Research on e-commerce transaction networks using multi-agent modelling and open application programming interface
- C7: Recommender (18 papers, confidence 0.665, PAJAIS: Recommender and Personalization Systems). Evidence titles: A semantics-based method for clustering of Chinese web search results | A neural network based reputation bootstrapping approach for service selection | A method for knowledge integration of ontology-based user profiles in personalised document retrieval systems
- C11: Control Manufacturing Process (18 papers, confidence 0.741, PAJAIS: IS Research Methods and Theory). Evidence titles: Linear approximation fuzzy model for fault detection in cyber-physical system for supply chain management | Cyber-Physical Systems, a new formal paradigm to model redundancy and resiliency | Data configurations in railway signalling engineering-an application of enterprise systems techniques
- C5: Method Algorithm Detection (16 papers, confidence 0.755, PAJAIS: Mobile Commerce and Applications). Evidence titles: A new architecture for simultaneous localization and mapping: an application of a planetary rover | Visual-inertial fusion SLAM system based on weighted mean filtering | Infrared target detection based on Gaussian model and Hungarian algorithm
- C18: Workflow Process Time (14 papers, confidence 0.712, PAJAIS: Mobile Commerce and Applications). Evidence titles: An approach to automatic development of interlocking logic based on statechart | Aligning observed and modelled behaviour based on workflow decomposition | An approach to repair Petri net-based process models with choice structures
- C22: Pricing Refactoring Wiki (6 papers, confidence 0.541, PAJAIS: Digital Transformation and Innovation). Evidence titles: Variable pricing business solutions in a decomposed business environment | Refactoring affordances in corporate wikis: a case for the use of mind maps | Architectural design for resilience

## Validation
Labels are validated through the in-app council table: keyword extraction, PAJAIS semantic mapping, and an LLM labeler when MISTRAL_API_KEY is configured. Without a key, the third council member uses a deterministic local semantic fallback, so the app remains executable end to end.

TCCM and computational technique extraction are exported in `tccm_validation.csv` for the top-cited 100 papers. Rows marked `NEEDS_FULL_TEXT_REVIEW` should be checked against PDFs before final academic submission. Full TCCM compliance requires uploading NotebookLM extraction and a second LLM extraction in the app's TCCM Dual Validation tab to generate `tccm_dual_validation.csv`.