Lets say we have certain products. We also have certain input features regarding these products like inventory, sales, price, cost etc. based on these features we want to score these parts. The score should reflect whether it is in shortage or not. Shortage is when there is high demand (Sales) and low inventory/supply. Score of 100 means it is in shortage. Score of 1 means it is not in shortage.
Right now i use a random forest model and use the probability from the model as the score. However this leads to a lot of products having score between 0-10 and 90-100. Very few products have score between 10-90.
Is there a better way to do this so that the scores are more distributed?