Help!- Churn for Bank Customers Analyzing

We have following dataset.

What is the best way to analyze the churn and how to get the max. precision??

We are new to Knime :frowning:

Hello @Highpredictor95 and welcome to the KNIME forum.

From my perspective, I don’t see here a KNIME question itself. I mean, if i focus straight to your question:

It depends on the purpose of the analysis, but at the end it’s an answer that you have to find by yourself, comparing the scoring of the different available Classification predictors available in KNIME; as accuracy depends as well on the data itself.

Navigating through the Node Repository Panel : Analytics > Mining you will be able to find all Classification Predictor nodes available in the platform. I would start from testing the Classification ones.

If you want to go beyond and research for the real maximum accuracy, then you should be able to compare KNIME extensions vs other open source data analysis tools.

Here you can find an analogue model Prediction of Customer Churn in a Bank Using Machine Learning / “The best answer to that question is: try them all!” using the same data you are pointing in your link.

BR

4 Likes

Hi @Highpredictor95 -

You may find this recent blog post on churn prediction in KNIME useful. Be sure to download and try out the example workflows!

4 Likes

We have following dataset.

We need the best possible prediction for Churn Rate and very detailed explanations because we are noobs.For the Knime Automation: We do not want a simple automation as you find it on google. We want to achieve max. accurancy and min. errorrate. Our usecase: We need to find the customers who are so valuable for us (the bank) that we contact them before they possible churn. Furthermore we need to examine what actions the bank can implement to prevent churn. In addition it would be great to learn something about, how much more it would cost to aquire new customers than to retain existing ones.

Is there somebody out there who can help me with this?