Skip to content

This repository consists of Rental Bike demand prediction required at each hour of the day so that stable supply of rental bikes can be made possible. This is done by applying various Regression Machine Learning Algorithms.

Notifications You must be signed in to change notification settings

SuhasTantri/Bike-Demand-Prediction

Repository files navigation

Rental Bike-Demand-Prediction

The goal of the project is to predict number of rental bikes required at each hour for stable supply of rental bikes. Various Regression machine learning algorithms have been applied on the dataset to get the best possible prediction. Some of the key takeaways from this project are -

  1. Performed Exploratory Data Analysis on the data to gain some insights.

2.Treated the outliers using log transformation.

  1. Label encoding was done for categorical variables.

4.Standard Scaler was used to scale down the data.

5.Applied 5 machine learning models on the dataset i.e,

  • Support Vector Machine.

  • K Nearest Neighbor.

  • Decision Tree.

  • Random Forest Classifier.

  • XGBoost.

6.Hyperparameter tuning was performed on all the models using Grid Search CV to get best parameters and best scores.

7.Random Forest Model performed best among the five models with train R2 score of 83.9 % and test R2 score of 82.23 %.

About

This repository consists of Rental Bike demand prediction required at each hour of the day so that stable supply of rental bikes can be made possible. This is done by applying various Regression Machine Learning Algorithms.

Topics

Resources

Stars

Watchers

Forks