Hi, I have a question, is it possible to make a recommendation system of which product to buy? A system similar to what Netflix or Amazon manages in their systems of recommending products?
For example, in my case, a sales system.
To recommend which products to sell to different customers according to what they buy.
To recommend, you need the recommendation system that detects:
if customer1 buys x things n times (and that buying pattern is repeated by n customers)
and customer 2 buys the same x product from customer1
then
The recommender system recommends to customer2 a product that customer2 could possibly buy according to the buying pattern of customer1 and n customers who make the same, similar purchases.
EXAMPLE TABLE
Products in inventory | customers | purchased product | times purchased | Recommended product |
|
|
|
|
|
Question.๐ค
What other variables are needed in order to make the recommender system?
I am very lost, I need help.
Have you checked Intelligence > Predictive Models, it might help.
Hmmmm.... ๐ค๐ค๐ค
A LOT would go into this, but starting small will reveal the way.
But to get started, you could do something like this:
The problem with predictive models is that they need to be individualized per person. You can't build a generalized predictive model based on the buying options of a group of people, the suggestions from the group probably won't align very well with what the individual would want.
User | Count |
---|---|
44 | |
31 | |
29 | |
14 | |
14 |