Skip to content

josephwilson8-nhs/AR_Miner

 
 

Repository files navigation

AR_Miner

Our implementation of the text miner for app reviews. Please see the original paper AR_Miner.

Getting Started

These python packages are required to run our project

sklearn

pip install sklearn

lda

pip install lda

scipy

pip install scipy

nltk

pip install nltk

nltk.corpus

Go to your python interactive model, type:

>>>nltk.download('all')

for install the nltk corpus to import stopwords

Run the project

Make sure to git clone the entire repo into your local directory and install the required packages.

Notebook structure

The notebook is used as a demo/report for our project. We import the code into the notebook.

The main code flow of our notebook includes :

  1. NLP Based Preprocessing;
  2. Naive Bayes/SVM Based Filtering;
  3. LDA topic clustering;
  4. Ranking algorithms for importance

Dataset

There are four considered apps to try:

facebook
templerun2
swiftkey
tapfish

About

The App Review Miner

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 94.5%
  • Python 5.5%