App Review Classification Model using Gated Recurrent Unit (GRU) and Flask Deployment

Overview

Gated Recurrent Unit (GRU) is an enhanced version of the standard Recurrent Neural Network (RNN) that addresses the problem of short-term memory and is well-suited for sequential data processing. This project focuses on building a review classification model using GRU to classify app reviews on a scale of 1 to 5, with 1 indicating negative sentiment and 5 indicating positive sentiment.

---

Aim

To build a classifier to classify app reviews on a scale of 1 to 5 using Gated Recurrent Unit (GRU).

---

Data Description

The dataset consists of app reviews and their corresponding ratings. The "score" column contains ratings ranging from 1 to 5 based on the "content" column.

---

Tech Stack

• Language: Python
• Libraries: pandas, TensorFlow, Matplotlib, scikit-learn, Pillow, Gunicorn, TextBlob, NLTK, Keras, Flask

---

Approach

Data Preprocessing

1. Converting words to lowercase.
2. Lemmatization of words.
3. Tokenization of words.
4. One-hot encoding of the scores.

Model Training

• Training a sequential model in TensorFlow.

Model Evaluation

• Evaluating the model on test data.

---

Modular Code Overview

1. Input: Contains the data used for analysis, including:

   • review_data.csv
   • test_review_data.csv (for testing)

2. Output: Contains the saved GRU model and pickle file.

3. src: The heart of the project, this folder contains all the modularized code for various steps. It further includes:

   • ML_pipeline: A folder with functions organized into different Python files, each appropriately named for its functionality. These Python functions are called inside the Engine.py file.

4. GRU-Neural-Network.ipynb and Model_Api.ipynb: The original Python notebooks.

5. requirements.txt: Lists all the required libraries and their versions for easy installation using pip.

---