Jakelee24 / Sentiment-Analysis-tensorflow

In this project I create the machine learning model which can classify postive and negtive result of the sentence by using word2vec and LSTM.

Sentiment-Analysis

Introduction

Sentiment analysis is contextual mining of text which identifies and extracts subjective information in source material, and helping a business to understand the social sentiment of their brand, product or service while monitoring online conversations.

In this project I create the machine learning model which can classify positive and negative result of the sentence by using word2vec and LSTM. enjoy it 🙂

Setup

Make sure you have the following is installed:

• Python3
• Tensorflow
• NumPy
• SciPy
• Pandas
• Matplotlib

Dataset

for the dataset please make sure your training data .txt file is like following:

Sentence	Class
I like this movie!	1
I hate this feeling.	0
...............	..

Pipline

The following is the project pipline:

Run

Notice: Before you start to run this project please make sure to analysis the vocabulary size of training data as following:

Once you prepared the training data, you can simply by execute the example.py to run the whole pipline.

python example.py

Or If you want to run the pipline step by step, please make sure you prepared the needed dataset for each module, Please check the Pipline chart to find it out.

python word2vec_module.py
python DataProcess_module.py
python LSTM_module.py