Skip to content

The data is related with direct marketing campaigns (phone calls) of a Portuguese banking institution. The classification goal is to predict if the client will subscribe a term deposit.

License

Notifications You must be signed in to change notification settings

Bharat-Reddy/Bank-Marketing-Analysis

Repository files navigation

Bank-Marketing-Analysis

  • Analyzed the prior marketing campaigns of a Portuguese Bank using various ML techniques like Logistic Regression, Random Forests,Decision Trees, Gradient Boosting and AdaBoost and predicted if the user will buy the Bank’s term deposit or not

  • Recommended, the marketing team, ways to better target customers using feature importance maps and business intuition

  • For More Information regarding dataset used, refer https://archive.ics.uci.edu/ml/datasets/Bank+Marketing

Instructions to run the code:

  1. All the data(raw data as well as pre-processed data) is present in "Data" folder.
  2. Make sure to run the notebook in python 3 environment. Make sure all the dependencies used in the notebook are installed in the local machine.
  3. The codes used for pre-processing the data is available in "Data Preparation" folder.
  4. The code used for Feature reduction using PCA is present in "PCA" folder.
  5. Finally the codes used for classification using different classification models is present in "Classifiers" folder.
  6. Notebook is commented adequately to give the rational, inferences of the executed code.

Quick result

Feature Distribution

Alt text

ROC curves of logistic regression model with features of degree 1, 2 and 3.

Alt text

Comparision of various classifiers.

Alt text

For detailed information regarding the project, please refer to the report.