1. Students’ Placement Office, IIT Kanpur
Project Verification Form
By appending your signatures to this form you acknowledge and agree that:
• This form along with the certificate would serve as the official document between the project supervisor and
Students Placement Office, IIT Kanpur regarding verification of the student’s project work
• The student will provide additional information and documentation relevant to his/her project upon request by
the Students’ Placement Office
• The student has clearly defined his/her individual role in projects done in cooperation with other students,
faculty, groups or company personnel.
• Incorrectly over-stating the reach, impact and/or quantitative/qualitative results of a project is unethical.
• In case of violation of any of the above rules, Students’ Placement Office, IIT Kanpur reserves the right to take
necessary action including de-registering the student from the placement season and reporting the misconduct
to the Institute Authorities.
Submitted by:- Project Supervisor Details:-
Name: Durga Kant Gupta Name: Mr. Somesh Kumar
Roll No: 13267 Designation: Vice President, Business Analytics,
IndiaMART Intermesh Limited
Title of the Project 1. Mapping existing products to their best Micro Category
2. Text mining to extract all the unique Brands & Specifications
3. Churn prediction of existing customers
4. Auto-Rejection of false buyer requirements
Commencement Date 9th
May 2016
Completion Date 8th
July 2016
Project Supervisor Mr. Somesh Kumar
Organization/Institution where the
Project was accomplished
IndiaMART Intermesh Limited
Head Office, Noida
Project Description (You can use extra A4 sheets in case you run out of space however the extra sheets should
also have the seal & signature of the Project Supervisor or the relevant authority )
Mapping of existing products to their most relevant Micro Category
• Developed, Analysed and initiated the Implementation of automatic Product Mapper which maps a
product to its most relevant Micro Category by considering the maximum string match( of product name
and MCAT) and maximum No. of leads for that Mcat in the previous 3 months data
• Compared this new algorithm with the existing search algorithm (live in system), it was found that on an
average 142 buyer requirements(leads) come to the Mcat which comes on top while being searched
• With new algorithm the no. of leads increased to 321, this gain of 179 leads per product will make
suppliers much more satisfied and engaged on IndiaMART portal
Text mining to extract all the unique Brands
• Created an extensive database of most recurred Brands for a particular Micro category of products using
Text Analytics so that they can form a separate Micro category
• Created another database of most frequent Specifications mentioned by buyers in the product
description to make only these specifications compulsory for the agent to enquire on phone and get rid-
off not so important ones which helped us to reduce the calling cost
Churn prediction of existing customers
• Created a model to predict whether an existing customer is going to renew his subscription at the end
of his subscription cycle or not
• Used decision tree C5.0 in R to create a large set of rules which were used for final predictions as stated
above

2. Signature: Signature:
Auto-Rejection of false positive buyer requirements
• Cost-Benefit analysis of auto-rejection of intent generated leads based on secondary Mcat match
• The exact no. of wrong buyer requirements which could have been auto-rejected was found to be
2690 while 8440 approved requirements will be auto-rejected by implementing the same
• After the conclusion that loss is more than the benefit, the idea of auto-rejection of leads based on
secondary Mcat match was not implemented