Skip to content

Project of Categorical Data Analysis 2020 fall in CUFE.

Notifications You must be signed in to change notification settings

Jace-Yang/bank-marketing-strategy-analysis

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

bank-marketing-strategy-analysis

Hi there! My name is Jace Yang. This was my final project of Categorical Data Analysis when I major Applied Statistics at Central University of Finance and Economics(CUFE) in 2020 fall.

  • code contains:

  • data contains:

    [Moro et al., 2011] S. Moro, R. Laureano and P. Cortez. Using Data Mining for Bank Direct Marketing: An Application of the CRISP-DM Methodology.

    • and some in-progress data stored in RData format to communicate with my teammates.
  • images and tables output used to write the report

Note: The Report is written in Chinese, but it contains some fancy charts I created, so you can still take a look. Also, I will elaborate on the may idea in the following README page.

Goal

  1. Forecast the success rate of the telemarketing calls, given information of target customers and the campaign records.

  2. Customer profiling for a effective marketing strategy.

Data Preparation

Data explore

Discrete variables

  • Personal Information:

  • Marital, Y Vs Job/Education/Housing:

Continuous variables

  • Account Information

  • Economics

Combined

  • Age vs job/education

Feature Engineering/Selecting

Lable Encoding:

  • Transform education type into education year:

Extracted features:

  • year features and monthly contact features from labelled month:

Delete/modify bad feature

  • weekday of last contact:

  • previous and poutcome are similar, so just keep one.

Missing value and ouliters handling:

We used Random Forest, Pmm linear prediction, mode/Average, and combined them.

Modelling

We used Logistics, SVM and decision tree.

Results: