Dynamic Attribute Weighting for Data Classification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current systems fail to dynamically adjust attribute weights based on new data records, leading to incomplete analysis and lack of visibility into spend patterns, particularly in complex procurement scenarios and email classification.

Innovation Solution

A computer-readable medium with instructions that analyze pre-classified data to assign static weights to attributes, calculate dynamic weights based on new record values, and apply classification algorithms like posterior probability and information gain to classify new records effectively.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If static attribute weights are used for classification, then the system is simple and fast, but the classification accuracy deteriorates when new data records arrive

Engineering Contradiction:

Improveclassification speedVSAvoidclassification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent implements dynamic attribute weighting where weights are no longer fixed but adapt automatically as new data records arrive. The system calculates information gain for each attribute based on new records and updates weights accordingly, allowing the classification model to evolve and improve accuracy over time without sacrificing computational efficiency

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system incorporates feedback mechanisms where new data records are continuously analyzed to update attribute weights. The information gain calculation provides feedback about which attributes are most informative for classification, and this feedback is used to adjust weights dynamically, creating a self-improving classification system

Inventive Principle:

Principle #23Feedback

2Measurement precision

If dynamic attribute weights are calculated for every new record, then classification accuracy improves, but computational complexity and processing time increase

Engineering Contradiction:

Improveclassification accuracyVSAvoidcomputational complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary calculations of information gain and weight updates in advance, preparing the classification model to handle new records efficiently. By pre-calculating attribute importance based on historical data patterns, the system reduces the computational burden when new records arrive, maintaining accuracy while managing complexity

Inventive Principle:

Principle #10Preliminary action

3Loss of information

If detailed spend analysis is performed across multiple dimensions, then visibility into spend patterns improves, but the system complexity and data processing requirements increase

Engineering Contradiction:

Improvespend visibilityVSAvoidsystem complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the spend analysis by dividing data into multiple dimensions (categories, suppliers, time periods) and applying classification algorithms to each segment. This segmentation allows detailed analysis of spend patterns across different dimensions while managing system complexity through modular processing of each segment independently

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS9009091B2Data classification tool using dynamic attribute weights and intervals of variation about static weights determined by conditional entropy of attribute descriptors

Publication Date: 2015.04.14 GLOBAL EPROCURE

AI summary

An analysis tool for causing a computer to use information gain of attributes and a classification algorithm to classify new records in a set of data by taking into account the predictive value of the attributes and the effect of the new record.