Predict the price of gifts

2.5

4 votes
Regression analysis, Machine Learning
Problem

Problem statement

One of the main goals of online retailers is to increase the desirability and the value of the products. To achieve this goal, various promotional techniques are planned. Among these techniques, offering promotions and special offers to the customers is an effective method of driving ancillary traffic to the site, acquiring new customers, and growing the revenue. These advancements can likewise be utilized to encourage new visitors to become loyal customers.

An online retailer has launched a special sale of gifts for the Good Friday event. They have approached you to build a model that helps them to set the prices for their gifts. The data provided is a combination of multiple categories and dates related to gifts.

However, based on these certain features you are required to build a model that can predict the prices of the different packages.

Data

The dataset consists of the following columns:

Data description

Column Description
gift_id Unique ID of gift
gift_type Type of gift (clothes/perfumes/etc.)
gift_category Category to which the gift belongs under that gift type
gift_cluster Type of industry the gift belongs
instock_date Date of arrival of stock
stock_update_date Date on which the stock was updated
lsg_1 - lsg_6 Anonymized variables related to gift
uk_date1, uk_date2 Buyer related dates
is_discounted Shows whether the discounted is applicable on the gift
volumes Number of packages bought
price The total price

The data folder consists of the following two .csv files:

  • train.csv - (20279×17)
  • test.csv - (13519×16)

The sample_submission is described as follows:

gift_id,price
GF_11156,175.54
GF_11157,90.4
GF_11158,102.0
GF_11159,130.05

Evaluation metric

score=max(0,100RMAE(actual_values,predicted_values))

Note: To avoid any discrepancies in the scoring, you must ensure that all the gift_id column values in the submitted file match the values in test.csv provided.

Click here to download the data.

Time Limit: 5
Memory Limit: 256
Source Limit:
Editor Image

?