Flair-Persian-NER Open-source Model - Free Deployment for Precise Recognition of 7 Categories of Persian Named Entities

Flair Persian NER

Developed by PooryaPiroozfar

Flair framework's built-in Persian 7-class named entity recognition model with an F1 score of 90.33 (NSURL-2019 dataset)

Sequence Labeling

PyTorch

Other#Persian NER #Multi-label Entity Recognition #High F1 Score

Downloads 3,846

Release Time : 1/9/2023

Model Overview

This model is used for named entity recognition in Persian text, capable of identifying 7 types of entities (person names, location names, organization names, dates, times, percentages, and monetary amounts). Built based on Flair embeddings and Pars-Bert model.

Model Features

Multi-category Entity Recognition

Can identify 7 different types of named entities, including person names, location names, organization names, etc.

High Accuracy

Achieves an F1 score of 90.33 on the NSURL-2019 dataset

Combination of Flair and Pars-Bert

Leverages the strengths of Flair embeddings and Pars-Bert model to improve recognition accuracy

Model Capabilities

Persian Text Processing

Named Entity Recognition

Sequence Labeling

Use Cases

Text Analysis

News Text Analysis

Extract key information such as person names and location names from Persian news

Accurately identifies various entities in news

Social Media Monitoring

Analyze entity information in Persian social media content

Helps understand key figures and locations in social media discussions

🚀 Persian NER Using Flair

This is a 7 - class Named - entity recognition model for Persian integrated with Flair. It aims to accurately identify named entities in Persian text, providing valuable information for natural language processing tasks.

F1 - Score: 90.33 (NSURL - 2019)

✨ Features

Predicted NER Tags

The model predicts the following NER tags:

tag	meaning
PER	person name
LOC	location name
ORG	organization name
DAT	date
TIM	time
PCT	percent
MON	Money

Model Basis

It is based on Flair embeddings and Pars - Bert.

📦 Installation

Requires: Flair (pip install flair)

💻 Usage Examples

Basic Usage

from flair.data import Sentence
from flair.models import SequenceTagger

# load tagger
tagger = SequenceTagger.load("PooryaPiroozfar/Flair-Persian-NER")

# make example sentence
sentence = Sentence("اولین نمایش این فیلم‌ها روز دوشنبه 13 اردیبهشت و از ساعت 21 در موزه سینماست.")

# predict NER tags
tagger.predict(sentence)

# print sentence
print(sentence)

# print predicted NER spans
print('The following NER tags are found:')
# iterate over entities and print
for entity in sentence.get_spans('ner'):
    print(entity)

This yields the following output:

Span[4:8]: "روز دوشنبه 13 اردیبهشت" → DAT (1.0)
Span[10:12]: "ساعت 21" → TIM (1.0)
Span[13:15]: "موزه سینماست" → LOC (0.9999)

📚 Documentation

Results

F - score (micro): 0.9033
F - score (macro): 0.8976
Accuracy: 0.8277

By class:
              precision    recall  f1 - score   support

         ORG     0.9016    0.8667    0.8838      1523
         LOC     0.9113    0.9305    0.9208      1425
         PER     0.9216    0.9322    0.9269      1224
         DAT     0.8623    0.7958    0.8277       480
         MON     0.9665    0.9558    0.9611       181
         PCT     0.9375    0.9740    0.9554        77
         TIM     0.8235    0.7925    0.8077        53

   micro avg     0.9081    0.8984    0.9033      4963
   macro avg     0.9035    0.8925    0.8976      4963
weighted avg     0.9076    0.8984    0.9028      4963
 samples avg     0.8277    0.8277    0.8277      4963

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご