Introducing 

Prezi AI.

Your new presentation assistant.

Refine, enhance, and tailor your content, source relevant images, and edit visuals quicker than ever before.

Loading…
Transcript

Analysis of Account tenure data

Naive Bayes Classifier

Data was classified into group_tenure: 0-12 years. 12-24 years, till > 60 years

Highest probabilities are in the tenure_group range from 0-12 years and 12-24 years

Trello Board

Correlation analysis

Customer status prediction for Oct Using Knn strategy

99.68%

Age 37 ~ 48: Visa + Mortgage + (Other.Loan) + Vehicle + Checking + Savings + Transactions

Variables in common: Visa + Vehicle + Checkings + Savings + Transactions

Age 49 ~ 60: Visa + (Mortgage) + Other.Loan Home.Equity + Vehicle + Checking + Savings + Transactions

99.75%

84.2%

Correlation analysis

99.1%

Age 13 ~ 24: Visa + Other.Loan + Vehicle + Checking + Savings + Transactions

Age 25 ~ 36: Visa + Mortgage + Other.Loan + Vehicle + Checking + Savings + Transactions

Predict next month's churn based on last month's median Bal. & transaction

Best Product

99%

200 active

Splited data into four groups based on account tenure, and then visualized balance for each product

Mortgage, Home equity, vehicle, CD, IRA, and Money Market are the products that active members use a lot and people tend to pull their money out when leaving the credit union

SVM

suppose it is August now, and we use July bal. to predict if customer will leave in September

Q1

Q2

Q3

Q4

100 active

Logistics regression

Boxplot for account tenure

Learn more about creating dynamic, engaging presentations with Prezi