I have a dataset of credit history of certain population i need to calculate the credit score for each one of them. What I'm planning is to calculate a probability on the basis of credit history variables then try to convert that probability into score. Is this approach gonna work or What technique or method should i follow?
First you will want to do some feature engineering in order to extract potentially meaningful indicators from your original dataset (unless it has been nicely preprocessed for you). Note that in order to do any ML you will need the credit score as a feature for each datapoint.
From there you will need to build a model. You will probably want to cast this as a regression problem with the credit score as the output. Most people usually start out with a simple linear model, this is a decent jumping off point: Regression.