This project attempts to answer the question posed by the Kaggle Bike Share question.
Clone the repo and save all files and directories in src/Rhtml/
to a place on your machine and open bike_share_report.html
in a browser (make sure the figure and static folders are in the same directory as the .html file).
The input data can be found in data/input
- test.csv The test dataset
- train.csv The train dataset
The output data (predicted values) can be found in data/output
- predicted values for all models
random_forest_counts.csv
contains the random forest regression predicted counts in the proper submission format
An Rhtml scipt src/Rhtml/bike_share_report.Rhtml
can be run using RStudio and knitr.
The R dependencies can be found in src/R/includes.R
- please make sure you install these
packages before running the Rhtml script.
Python dependencies include having Python3 installed and the following packages installed:
- scikit-learn
- numpy
- pandas
- statsmodels
Make sure necessary execute permissions have been added to the python script src/python/machine_learning_algorithm.py
. From the root directory bike_share
,
cd src/python/
python3 ./machine_learning_algorithm.py
This will run the following models:
- Linear Regression
- Poisson Regression
- Lasso Regression
- Ridge Regression
- Random Forest Regression
Note that you can change the constant values in that file to test models with different arguments and features.