What do i need to create a system that classifies bank packages for customers using previous customer data in python?



I want to develop a system that classifies bank packages for customers using neural networks and big data. How do I tackle this? I want to use previous customer data like spending behavior, salary, type of employment, age etc. to classify the customers into bank packages e.g Gold, Silver, Bronze.


@maffsojah, why neural networks? Do you have any specific reason to use it? Have you tried simpler methods, like logistic regression or trees? Seriously, you don’t want to mess with NNs unless you are fairly certain that you can benefit from them. Spending a lot (and I mean A LOT) of time and resources training and tuning NNs just to find out that clustering + logistic regression or something do almost as good a job is terrible. And I assure you it happens.

So, if you haven’t tried any of these, please do. Engineering new features is also critical.

In case you already tried the simpler solution, crafted some features and are still not satisfied, you may read this introduction to NNs with TensorFlow.


Thanks @caiotaniguchi. The requirements say I need to use NNs for the classification


Hi maffojah,

If data is huge in size and your single machine is not able to process it then you should try Bigdata frameworks, like Apache Spark/ Hadoop.

Here I am sharing you a link which introduce Apache Spark and a machine learning technique to solve the real world problem.



Thanks @facebook_user_4, I have been reading on apache spark and for a while it seemed like It would solve this. Now I cannot find any help regarding multi-label classification using pyspark.