I'm trying to find a way to predict an integer value based off of an item's prior sale history and am inquiring for a starting point on how to approach this.
The factors are as follows (Data Type in Parenthesis):
A) ReportDate (DateTime)
B) Part# (String)
C) Part Type (String)
D) Lost Sales in past 150 days (Int)
E) Sales in past 150 days (Int)
F) Quantity currently in stock (Int)
G) Quantity to be Purchased (Int)
My goal is to minimize Factor D so that it closes in at 0 as Factor G has better inputs. It should also account for what is already in stock aka Factor F
Factor G is manually inputted data while the other data is automatically generated. Factor G's value is based on the item's sale history.
There is a monthly reporting of each item spanning over the past 5 years.
For now, I am just concerned with being able to predict one item correctly.
So if I had 5 years of data from 2013-2018, that would be 60 rows of data.
It is now January 2019, so I should expect a row of data with Factor G missing, this is where I would need to predict a value.
Also, is it possible to account for report dates since Lost 150 and Sales in past 150 are dependent on the past 5 rows?
