Uplifting Model

A retailer has collected some data from its historical promotions outcomes in terms of customers, purchases, recency (last time customer purchased), and conversion to promotions (if the customer has purchased or not). With this historical information, he wants to know which customers are more likely to purchase and what products. Machine learning can learn a model relating these collected features to customer conversion. They can use this model as a tool for planning their next promotional campaigns. This retailer has collected the following data:

• recency: how recently a customer has made a purchase
• history: sum of all purchases
• used_discount: if customer have used a discount campaign before
• used_bogo: if customer have used a bogo (“buy one get one”) campaign before
• zip_code: some demographical information
• is_referral: referring someone for some product
• channel: purchase channel
• offer: Three options [“No offer”, “Discount” os “Buy one get one”]
• conversion: Two options [0: No, 1:Yes] TARGET