Project Setup: BikeSharingDemand Jupyter notebook is attached and hence it must be downloaded by the user to run the project.
Problem Statement: predict the total count of bikes rented during each hour
Description: Using the trained model for forecasting the bike demand using the test data set.
Procedure: EDA - Exploratory Data Analysis, to visualize the important features in the data set, to find some outliers, also to find correlation between the different features in the data set. 80-20 split of the train data into training, and validation to predict bike rental demand. Maximum likelihood estimation in Poisson regression was carried out using Gradient Descent Approach and parameters were estimated. L1 and L2 norm regularization over weight vectors, were used to find the best hyper-parameter settings .
The accuracy on test data for no regularization, L1 norm regularization and L2 norm regularization were reported.
Regression Used for Prediction: Poisson regression was used for predictions. Coding Language: Python is used as a programming language. Libraries used: 1.Pandas 2. Numpy 3.Matplotlib 4.Seaborn Live code, visualizations and narrative text: Jupyter notebook is used for live coding and for proper visualizations with the help of graphs.