Persian Text Analysis Projects

This repository contains two Persian text analysis projects: Sentence Tokenization and Sentiment Analysis.

Project 1: Persian Sentence Tokenizing

This project tokenizes sentences in Persian texts. It splits the input text into individual sentences and labels them based on the given ratings.

Features

• Sentence Tokenization: Splits the text into separate sentences.
• Sentiment Labeling: Labels sentences based on their rating as positive, neutral, or negative.

Usage

1. Install Required Libraries
2. It's suggested to use Python 3.6 or 3.7 for sentence tokenizing and Parsinorm project.
3. It's suggested to run the tokenizing project in a conda environment.
4. Please visit the link and add any necessary files to the Anaconda directories.
5. It's recommended to use Python 3.11 for running sentiment analysis.

Run the Script

Execute the script using Python. The script will prompt you for the following inputs:

• File Path: Path to the input CSV file.
• Text Column Index: Index of the column containing text data (e.g., 0 for the first column).
• Rating Column Index: Index of the column containing rating data (e.g., 1 for the second column).
• Output File Name: Name of the output CSV file where the tokenized data will be saved (e.g., tokenized_output.csv).

The script will read the CSV file, tokenize sentences, and save the results to the specified output file.

Project 2: Sentiment Analysis

This project performs sentiment analysis on Persian sentences using the pre-trained "roberta-sentiment-persian" model. It classifies sentences into one of three categories: negative, positive, or neutral.

Features

• Sentiment Analysis: Classifies sentences into negative, positive, or neutral.
• Pre-trained Model: Uses the "roberta-sentiment-persian" model for sentiment classification.

Usage

1. Install Required Libraries

   Install the necessary libraries using the following command:

   pip install transformers torch pandas
   

Run the Script

Execute the script using Python. The script will prompt you for the following inputs:

• File Path: Path to the input CSV file.
• Text Column Index: Index of the column containing text data (e.g., 0 for the first column).
• Output File Name: Name of the output CSV file where the sentiment labels will be saved (e.g., sentiment_output.csv).

The script will load the model, analyze the sentiment of each sentence, and save the results to the specified output file.