/Foretelling-Retail-price-using-Multivariate-Regression

Primary LanguageJupyter NotebookMIT LicenseMIT

Foretelling-Retail-price-using-Multivariate-Regression

The clients wanted to find if the prices they were being charged / produced invoice for, were reasonable or were they being taken advantage of.

DATA

Data provided is divided into TRAIN SET and TEST SET. Training data has $284780$ observations and $8$ variables (including target), and Test data has $122049$ observations and $7$ variables.

Attribute Description:

Attribute Description
Invoice No Invoice ID, encoded as Label
StockCode Unique code per stock, encoded as Label
Description The Description, encoded as Label
Quantity Quantity purchased
InvoiceDate Date of purchase
UnitPrice The target value, price of every product
CustomerID Unique Identifier for every country
Country Country of sales, encoded as Label

Machine Learning Formulation

The use case lends itself as a Multivariate Regression problem.

Performance Metric

RMSE (Root Mean Square Error)