Co-authored by Alex Jorenby and Jiehwan Yang
Description
When a seller inputs an item's information (category name, brand name, item condition, shipping, and item description), Mercari wants to suggest an ideal selling price so that the seller can make more informed decisions with setting a price for each item.
To solve this problem, we used natural language processing (NLP) techniques on the item descriptions (145 characters on average per item) and used regression algorithms to predict the item’s price.
We utilized Docker and Git to collaborate.
Lessons Learned
It was more than a machine learning project I've worked on in the past.
It was an ML project with "collaboration".
Docker
We were surprised by its power to make collaboration easier by enabling the application to run in different environments, so that we could run each other's work on our local machines without having to worry about syncing up system settings and environments.
Git
I have used Git to build a GitHub blog page before, but it was my first time using Git to collaborate with someone else. I hope this collaborating experience with Git will help me when I start working in industries.
I am sure there is a lot more to explore in both of these collaboration tools, but I am still glad that I got a hang of them!
File source:
https://github.com/jiehwan94/Price_Prdiction_Using_NLP
Kommentare