How to use Data Mining to automatically set the selling price?

Question

data mining

Good afternoon!
I was faced with the task of reducing time costs when setting prices for goods that arrived in electronic waybills.
There is a database of statistics on the set prices. It is necessary, using Data Mining algorithms, to obtain the relationship between the "Selling Price" and a set of parameters on which the price depends - "Supply Price", "Previous Selling Price" and 2 more parameters. Thus, providing automatic setting of prices with the possibility of adjusting the algorithm, when business conditions change.
Unfortunately, the first attempt to create such an algorithm failed. I tried to write the classification rules myself, according to the words of the specialist putting down the prices, for the entire product range.
The result of the first attempt to build an algorithm showed that there are several heterogeneous groups of goods in the commodity turnover, which are evaluated according to different criteria.
Since I am not a specialist in Data Mining, I am just learning, I would like to know:
1. Is it necessary, using cluster analysis, to divide the product range into 8-9 groups and then for each group separately, using the regression apparatus, find the relationship between " Realization price" and a set of parameters. Or will the classification algorithms themselves (for example, Decision Trees) when adding the product code parameter, themselves, based on the initial data, create the necessary groups in the tree?
2. What is the best regression algorithm to use, provided that most of the parameters are not categorical variables, but numerical ones?
3. Which Data Mining package is better to use (excluding commercial products), provided that the data for analysis is in MS SQL DBMS?
4. List of independent parameters:
4.1 Delivery price;
4.2 Previous selling price;
4.3 Vital and essential drug sign (Vital and essential medicine - the list of medicines is approved by the Government of the Russian Federation. For such medicines, the price cannot exceed a certain percentage of the maximum allowable price);
4.4 Price according to the State Register for vital drugs;
4.5 Product code (I'm not sure what is needed, since the volume of the product range is about 31 thousand)
4.6 Product group code (I'm not sure, because groups do not exist, but are only planned to be created)
Best regards,
Alexander.

Answer 1

If everything is in MS SQL, then right in it for each product group or product you need to build a regression
Description here https://msdn.microsoft.com/en-us/library/cc645906.aspx