Phrase-Based Data Classification System for Job Titles

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing data classification systems, such as SVMs, are not effective in classifying self-descriptive job titles due to the variability in how users specify their job titles, leading to low precision and coverage, especially when dealing with large datasets like those on social networking sites.

Innovation Solution

A phrase-based classification system that employs lazy learning from labeled phrases, utilizing the Information Bottleneck principle to quantify the near-sufficiency property, allowing for efficient classification by creating a controlled vocabulary and building phrases based on text data items, which significantly outperforms predecessor systems in precision and coverage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If standard classifiers (e.g., SVMs) are used to classify job titles, then the classification process can be automated, but the precision and coverage are low due to variability in user-specified job titles

Engineering Contradiction:

Improveautomation of job title classificationVSAvoidprecision of job title classification

Core Design Contradiction:

Extent of automationVSMeasurement precision

Solution Approach 1:

The patent introduces an intermediary controlled vocabulary layer between the variable user input and the classification algorithm. User-specified job titles are first mapped to standardized phrases from the controlled vocabulary, which then serve as the actual classification features. This intermediary structure resolves the contradiction by maintaining automation while improving precision through standardized representation.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transforms the classification parameters by changing from raw user-specified job titles to standardized phrases from a controlled vocabulary. This parameter transformation allows the same classification algorithm to achieve higher precision by operating on normalized, standardized features rather than variable natural language input.

Inventive Principle:

Principle #35Parameter changes

2Ease of manufacture

If standard classifiers are used for job title classification, then implementation is straightforward, but coverage is low when dealing with 40,000 different ways users specify the same job category

Engineering Contradiction:

Improveease of implementing classification systemVSAvoidcoverage of job title classification

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent performs preliminary action by pre-defining a controlled vocabulary of standardized job title phrases before the classification process. This controlled vocabulary serves as a pre-prepared mapping structure that covers 40,000+ user variations. By preparing this vocabulary in advance, the system achieves high coverage without complicating the actual classification implementation.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If a controlled vocabulary and phrase-based approach is used, then precision and coverage improve, but the system complexity increases

Engineering Contradiction:

Improveprecision and coverage of classificationVSAvoidcomplexity of classification system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the classification system into distinct modular components: (1) controlled vocabulary management, (2) phrase building from vocabulary, (3) phrase-based classification. This segmentation allows each component to be independently managed and optimized, reducing overall system complexity while maintaining high precision through specialized handling at each stage.

Inventive Principle:

Principle #1Segmentation

4Measurement precision

If manual labeling is used to create the controlled vocabulary, then classification quality improves, but the time and resources required increase

Engineering Contradiction:

Improvequality of classificationVSAvoidtime for manual labeling process

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by manually labeling only the essential controlled vocabulary phrases rather than attempting to manually label all possible job title variations. The controlled vocabulary covers the core categories, and the phrase-building mechanism automatically handles the remaining variations, reducing manual labeling time while maintaining high classification quality.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS9268766B2Phrase-based data classification system

Publication Date: 2016.02.23 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

A method of classifying data is disclosed. Text data items are received. A set of classes into which the text data items are to be classified is received. A phrase-based classifier to classify the text data items into the set of classes is selected. The phrase-based classifier is applied to classify the text data items into the classes. Here, the applying includes creating a controlled vocabulary pertaining to classifying the text data items into the set of classes, building phrases based on the text data items and the controlled vocabulary, and classifying the text data items into the set of classes based on the phrases.