Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this community
We have data consist of previous transaction history consisting of Date,Order-id, Product-id, Product name, ordered or not. We need to predict a specific product probability for all the customers that will buy or not
I am a beginner and having difficulty in coming up with a model. Any input is highly appreciated.
Thank You
There don't seem to be many variables involved. Since you're asking for the probability of customers buying a product, You could use a simple logistic regression model. It's used for classification using the sigmoid curve. It will give you a cutoff probability, above which you can consider customer would buy a product.
When you are providing only one option (will it sell or not). This is assuming there are no consumer choices that may not be a correct assumption. You're model is more closely related to a coin flip. You have one choice (you have no choice).
If the question is closer to, "Given two or more products what is the likelihood a consumer would purchase A over B (over C (and on)).
When you ask this you are looking at receipts, historical (assuming you have enough to go on) and you are also building some simple A/B testing against your mystery product.
What model will be apt for this scenario?
Your potential models for A/B testing could be: The frequentist approach or Bayesian statistics approach.
What features we need to consider?
Historical data. Consumer Options. A/B testing (a hypothesis).
