Datasets:

U4RASD
/

Masrad

dataset_info:
  features:
    - name: semantic
      dtype: float64
    - name: semantic_rank
      dtype: float64
    - name: semantic_diff
      dtype: float64
    - name: lexical1
      dtype: float64
    - name: lexical1_rank
      dtype: float64
    - name: lexical1_diff
      dtype: float64
    - name: lexical2
      dtype: float64
    - name: lexical2_rank
      dtype: float64
    - name: lexical2_diff
      dtype: float64
    - name: entity
      dtype: string
    - name: src_entity
      dtype: string
    - name: phonetic
      dtype: bool
    - name: pos
      dtype: string
    - name: label
      dtype: bool
  splits:
    - name: train
      num_examples: 19405
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train.csv
license: mit
task_categories:
  - text-classification
language:
  - ar
size_categories:
  - 10K<n<100K

Masrad Dataset

Dataset Description

The Masrad dataset contains 19,405 examples with features related to semantic similarity, lexical similarity, entity recognition, phonetic matching, and part-of-speech tagging for Arabic text.

Features

Feature	Type	Description
`semantic`	float	Semantic similarity score
`semantic_rank`	float	Rank based on semantic similarity
`semantic_diff`	float	Difference from top semantic score
`lexical1`	float	First lexical similarity score
`lexical1_rank`	float	Rank based on first lexical score
`lexical1_diff`	float	Difference from top first lexical score
`lexical2`	float	Second lexical similarity score
`lexical2_rank`	float	Rank based on second lexical score
`lexical2_diff`	float	Difference from top second lexical score
`entity`	string	Detected entity type (PER, ORG, LOC, none)
`src_entity`	string	Source entity type
`phonetic`	bool	Whether phonetic match exists
`pos`	string	Part of speech tag
`label`	bool	Target label

Usage

from datasets import load_dataset

dataset = load_dataset("U4RASD/Masrad")