CS587Project

Title: Price Strategy Optimization For Online Sales Regarding Fake Review Detection

Questions

1. Can fake reviews detected only by looking through text and ratings?

code and analysis are located in phase2.ipynb or phase2.pdf. More specifically, section Task1: classification of fake and original reviews is related to this question. This section contains code blocks associated with this question and markdown blocks containing explanation, analysis, and discussion.

2. Do factors like the probability of being a fake review, ratings, etc. impact prices?

code and analysis are located in phase2.ipynb or phase2.pdf. More specifically, section Task2: Price regression analysis is related to this question. This section contains code blocks related to this question and markdown blocks containing explanation, analysis, and discussion.

Folder structure

this project is organized in this way:

.
├── app               # folder containing app and database codes
├── exp               # folder containing all ipynb files from previous phases (1&2)
├── data              # folder containing datasets csv files
├── demo_video.mp4    # demo video of making the app and its features
└── cs87 report.pdf   # report PDF file

Buliding the app and database

1. Install the required libraries:

pip install scikit-learn numpy pandas plotly matplotlib contractions nltk bs4 xgboost streamlit

2. change the directory to this project folder and app foler:

cd PROJECT_DIR/CS587Project/app

3. For making the database you can run this code or simply use the db file made before in the app folder:

python db_creation.py

4. Make the app by executing this line:

streamlit run app.py