Instillation Instructions

Clone the repository

Make a directory

git clone https://github.com/EECS486/Bezos.git

Install Virtual Env

Virtual Env acts as a virtual enviornment so that we can virtually install python packages and not overwrite the ones on our system

pip install virtualenv

Create Virtual Env and Enter it

Go to your directory where you cloned the repository

$ cd Bezos

$ virtualenv -p python3 env

$ source env/bin/activate

Install the requirements

• Make sure in home directory

pip install -r requirements.txt

Extra: If You Want to Install A New Package

$ source bin/activate

pip install package

pip freeze > requirements.txt

$ deactivate

File + Folder Descriptions

1. data_analytics_erneh.py - displays analytics for the review and metadata data

2. jsonReviewRead.py - review parser for classification models

3. jsonReviewRead_vBlackfyre.py - review parser for classification models

4. metadata.py - metadata parser for classification models

5. naivebayes.py - naivebayes model classifer

6. porter.py - porter stemmer

7. reviewdata.py - reviewer parser for naive bayes model

8. reviewdataNB.py - review parser for naive bayes model

9. linking_and_metrics.py - links the metadata and review data and calculates analytics for each review

10. modelGeneration.py - generates each classification model and projects helpfulness of amazon reviews

11. output/ - output for naive bayes and classification models

12. plots/ - plots for feature importance

Instructions

1. Download and Extract the contents of this folder into the repository https://drive.google.com/file/d/1QCZXLE9F9BqI2k2y3APi4tPItdREnhMS/view?usp=sharing

• It contains the json review and metadata files used to generate the review data along with pickle files to make the model generation faster

Naive Bayes

1. Open Naive Bayes
2. set the line: params = {"stem": False, "stop": True, "condProb": True, 'bigram': True} to what parameters you want to run the naive bayes with

• Stem: stems words
• stop: remove stop words
• condProb: make sure True
• bigram: creates a bigram model instead of a unigram model

Classification Models

1. Install Stanford CoreNLP https://stanfordnlp.github.io/CoreNLP/index.html
2. Go To Directory Where Installed Stanford CoreNLP
3. Follow Instructions to Run Stanford CoreNLP on Port 9000 https://stanfordnlp.github.io/CoreNLP/corenlp-server.html#getting-started
4. In linking_and_metrics.py

• Set category to the category you want to process. ie: 'Grocery_and_Gourmet_Food'

5. Run data_analytics_erneh.py to generate statics on the whole review set
6. In modelGeneration.py

• Set category to the category you want to process. ie: 'Grocery_and_Gourmet_Food'
• Note: pkl files must be generated for said category