Job Matching System Using Weighted Semantic Scoring

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current job search tools often require users to sift through numerous results or return no results if criteria are too specific or narrow, failing to effectively match job seekers with suitable job openings.

Innovation Solution

A job matching search tool and method that processes job-descriptor data to derive and store words, root forms, and synonyms, allowing users to classify job openings using swipe gestures, and ranks matches based on generated scores for intelligent job recommendations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional keyword-based searching tools are used to match job seekers with job openings, then the search process becomes simple to operate, but the matching accuracy deteriorates resulting in irrelevant results or no results when criteria are specific

Engineering Contradiction:
Improvematching accuracyVSAvoidsearch tool complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent transforms the matching parameters from simple keyword presence/absence to weighted semantic similarity scores. Job descriptors and candidate profiles are converted into weighted word lists where each word has an associated importance weight. The matching process calculates a composite score based on weighted overlaps between job requirements and candidate qualifications, enabling nuanced differentiation among candidates and jobs while maintaining operational simplicity through automated scoring.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If users manually review and analyze job descriptions to determine matches, then the matching accuracy improves, but the time required for the search process increases

Engineering Contradiction:
Improvematching accuracyVSAvoidsearch time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary processing of job descriptions and candidate profiles before the actual matching occurs. Job descriptors are pre-analyzed to extract weighted keywords and phrases that capture essential job requirements. Candidate profiles are similarly pre-processed to identify relevant qualifications. This preliminary structuring enables rapid automated matching without requiring manual review, significantly reducing search time while maintaining high matching accuracy through the pre-established weighted parameter frameworks.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If specific criteria are provided in the search to improve matching precision, then the relevance of results improves, but the quantity of search results decreases potentially to zero

Engineering Contradiction:
Improveresult relevanceVSAvoidnumber of search results
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The matching system evaluates all job descriptors and candidate profile elements, not just exact keyword matches. It calculates partial scores for each weighted word overlap, allowing candidates to receive positive match scores even if they don't meet every single criterion perfectly. The system ranks results by cumulative weighted scores, ensuring that a sufficient quantity of potentially relevant jobs are returned even when specific criteria are provided, while still maintaining high relevance through the weighted scoring mechanism.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS11176152B2Job matching method and system
Publication Date: 2021.11.16 TORRE LABS INC
  • US11176152B2 patent drawing
  • US11176152B2 patent drawing
  • US11176152B2 patent drawing

AI summary

A computer-implemented method and system is provided that involves storing job-descriptor data for a plurality of job openings and processing the job descriptor data to derive and store words or root forms or synonyms corresponding thereto. A job seeker user interacts with the system to review at least one job opening, which involves the user classifying at least one word or text phrase of the job-descriptor data for the at least one job opening, deriving and storing words or root forms or synonym corresponding to the at least one word or text phrase classified by the user, and ranking a set of job openings based upon scores for the set of job openings, where the scores are based on frequencies of the at least one word or text phrase classified by the user. Other features and aspects are described and claimed.