Can you predict product backorder?

Problem Statement

Material backorder is a common problem in a supply chain system, impacting an inventory system's service level and effectiveness. Identifying parts with the highest chances of shortage prior to their occurrence can present a high opportunity to improve an overall company’s performance. In this project, we will train classifiers to predict future back-ordered products and generate predictions for a test set.

File descriptions

Here we have two CSV files (Training_BOP.csv and Testing_BOP.csv)

Training_BOP.csv - the training set

Testing_BOP.csv - the testing set

Each file has 23 columns; the last column (went_on_backorder) is the target column.

Data fields

sku - sku code

national_inv - Current inventory level of component

lead_time - Transit time

in_transit_qty - Quantity in transit

forecast_x_month - Forecast sales for the net 3, 6, and 9 months

sales_x_month - Sales quantity for the prior 1, 3, 6, and 9 months

min_bank - Minimum recommended amount in stock

potential_issue - Indicator variable noting a potential issue with the item

pieces_past_due - Parts overdue from the source

perf_x_months_avg - Source performance in the last 6 and 12 months

local_bo_qty - Amount of stock orders overdue

x17-x22 - General Risk Flags

went_on_back_order - Product went on backorder

Validation - indicator variable for training (0), validation (1), and test set (2)