sentiment-analysis

AI based Context and Sentiment Analysis for Messaging

Prerequisites

• Create a Telegram Developer App by using this guide https://core.telegram.org/api/obtaining_api_id
• Add the values for AppId, AppHash, Usernamem, and PhoneNumber in TelegramClient.ipynb

To train the model

• TrainModel.ipynb: Used to train models and then save them inside 'results' folder

• To train model on different datasets, you can change the value of 'Data' variable, and can select any one value from these:

  1. formspring
  2. wiki
  3. twitter
  4. sentiment140

  By default training is done on Wikipedia datasets

To run the Project

• TelegramClient.ipynb: Connects to the Telegram application, reads the chats and predicts the sentiment using the trained model
• Run and authenticate with the your user and telegram app info
• To find out the sentiment score, send "<score>" message in telegram