data-mining-proj

Group 31, Chuan De Sheng | Jordan Lam| Nathanael S Raj | Ng Tze Yang | Phang Jun Yu | Sarah Ye

Instructions to Set Up

1. Make sure you have Anaconda (latest) installed
2. For existing user, run conda update conda, then conda update --all to make sure the latest packages are installed
3. On your conda prompt, type: conda create --name mining --file dependencies.txt . If you desire to run python3.6, use: conda create --name mining python=3.6 --file dependencies.txt
4. Navigate to the project directory, and activate the environment For Windows: activate mining For Mac/Linux: source activate mining
5. Run python main.py to run all classifiers. Results are printed onto terminals and graphs are produced.

Good Project Practices

1. In the main code (ie main.py) there is a sandbox area for you to write the code that does your part.
2. After writing that, copy and paste the code as a function into your own python module.
3. After that, in main.py, import the python module function and store the output in a variable so that the next person down the pipeline can easily access your data if needed.
4. If you need an example, you can look at the importcsv.py and processdata.py packages, and how I implemented those functions as a one liner in the main.py code
5. Pull before pushing, or create your own branch to ensure no code clashes exist! It is a pain in the ass to resolve merge errors.
6. If you ever import a new dependency, remember to import it using conda e.g. conda install PACKAGE_NAME. After that, update the dependency using conda list --explicit > dependencies.txt
7. If there is a dependency change, key in the command conda install --yes --file dependencies.txt while in the mining environment

Have a good time :)

Notes on creating machine learning functions

1. For each classifier you are creating, e.g. svm, random forest etc., please create a predict function that takes in testX and returns the corresponding predictions
2. For example, create a function bayesPredictions = bayesian.naiveBayes(testX, testY, trainX, trainY) where bayesPredictions will represent the classifications of testX.
3. Place this function in main.py
4. You can look at the current code in main.py to get an idea