What is collaborative filtering method

Dołączył: 11 Mar 2024 Posty: 1

Ishikura: What Workship has introduced this time is a method called "collaborative filtering." I will explain this in detail. 1. Find users similar to the target user (similarity calculation, correlation analysis) Ishikura: I'll replace it with probability. 2. Calculate frequently viewed jobs based on similar users and decide which jobs to recommend to that user. Ishikura: However, the problem with this method was that the more users there were, the more the amount of calculations would increase by the multiplier.

The solution was to create an INDEX so that all users did not have to view it. I was deeply moved by the amount of effort that went into creating the job posting websites that I use casually. Recommendation Chinese Student Phone Number List engine improvement proposal Next on stage was backend engineer Koki Sakamoto. He introduced ways to improve Workship's recommendation accuracy. Koki Sakamoto: Backend engineer. During his student days, he conducted network research related to brain science and researched and developed speech recognition systems using deep learning.

He joined GIG in October 2018 and belongs to the SaaS business department as a backend engineer. He has improved the system that is the basis of contracted development, contributing to increased productivity. Here are the slides used that day. LT13 (second half) Recommendation engine implementation in Workship from GIG inc. Improvement idea 1. Replace with “0, x” instead of “0 or 1” Sakamoto: The first improvement idea is to change the value to multiple discrete values or a continuous value with a range, rather than a binary value of 0 or 1. By expanding the space in this way, it becomes easier to draw a line between the presence and absence of a correlation coefficient.

_________________
Chinese Student Phone Number List