In this repository, CPC Bidding task is implemented.
CPC Bidding (Cost-per-click) is a charging metric for (booking, advertisement, etc.) websites.
In this project, a booking website's 90 days of data is preprocessed, analysed and a forecasting model is proposed. Lastly, the model's performance is evaluated.
A detailed report can be found in here.
Source code of the report can be found in here.
Code is implemented in Python 3.5 on Jupyter-Notebook.
General Functions: subprocess, warnings, os
Data Manipulation: pandas, numpy
Data Visualization: seaborn, matplotlib
Data Preparation: sklearn.utils, sklearn.model_selection
Machine Learning Model Proposal: sklearn.ensemble, sklearn.multioutput
Performance Evaluation: sklearn.metrics
In the report, the whole process of the solution is expressed.
- Walking through process
- Data Preprocessing
- Data Analysis
- Feature Selection
- Modeling
- Performance Evaluation
- Conclusion