When a new product is about to be introduced to the market, is often convenient to create a sale forecast and estimate the quantity of the good that can be sold over a defined period of time. This could be a very useful piece of information for the manufacturer, or the seller in general, in order to come up with advantageous insights and to make correct decision about the quantity to be produced/put on the market or the right price to assign to the product. Knowing if a product is going to be sold out everywhere immediately after its release or if instead very few people would want to buy it, can be beneficial also to the customer, who can decide to wait and make the purchase at a lower price or simply avoid long lines at the stores.
With this project I tried to model with a Bayesian Network the behavior of the market with respect to the launch of a new product and in particular to predict its availability at the stores after its release. Bayesian Network allowed me to deal with uncertainty and partial information, and to make probabilistic predictions by defining probabilities of different causes that could affect my target in a very simplified model of my domain of interest.
For a more detailed discussion about this project read the report
Packages requirements
pgmpy 0.1.14