Machine Learning Model for Standardized Component Recommendations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional management systems require unique configurations for each entity providing services and components, making it inefficient to generate standardized recommendations across multiple entities due to the use of different identifiers for similar services and components.

Innovation Solution

A management system standardizes service and component identifiers by generating a training dataset through categorization of historical entries into predetermined classifications, allowing a single machine learning model to provide recommendations across multiple entities by predicting service and component likelihoods based on user attributes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional management systems use entity-specific identifiers for services and components, then each entity can maintain its own unique identification system, but the system complexity increases and scalability decreases when deploying across multiple entities

Engineering Contradiction:

Improveadaptability to different entity identifier systemsVSAvoidsystem configuration complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces a machine learning model as an intermediary layer between entity-specific identifiers and standardized service/component classifications. The ML model learns mappings from various entity identifier systems to a universal classification schema, enabling the management system to handle multiple entity-specific identifier systems without requiring separate configurations for each entity. This intermediary translates diverse input formats into a standardized internal representation.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent creates a universal service and component classification schema that can serve multiple entities with different identifier systems. Instead of configuring the system separately for each entity, a single ML model is trained on aggregated data from multiple entities and deployed universally. This universal model handles the adaptation to different entity-specific identifiers while maintaining consistent service recommendations across all entities.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Reliability

If a management system is uniquely configured for each entity, then recommendations can be tailored to entity-specific identifiers, but the deployment time and resource requirements increase significantly

Engineering Contradiction:

Improverecommendation accuracyVSAvoidmodel training and deployment time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent merges training data from multiple entities into a single aggregated dataset. Instead of training separate ML models for each entity, the system combines historical service and component data from all entities, using entity-specific identifiers as input features. This unified training approach reduces the total number of models required and decreases overall deployment time while maintaining recommendation accuracy through the ML model's ability to learn entity-specific patterns from the aggregated data.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent performs preliminary aggregation and preprocessing of data from multiple entities before model training. By consolidating data upfront and creating a unified training dataset that includes various entity-specific identifier formats, the system prepares the foundation for a single universal model. This preliminary action eliminates the need for repeated data collection and preprocessing steps that would be required if training separate models for each entity.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If historical service entries from different entities are used to train a single ML model, then scalability improves and deployment becomes easier, but the difficulty of data standardization and categorization increases

Engineering Contradiction:

Improverecommendation generation efficiencyVSAvoiddata categorization complexity

Core Design Contradiction:

ProductivityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent employs a dynamic approach to data categorization by using the ML model to learn mappings from diverse entity-specific identifiers to standardized service and component classifications. Instead of relying on rigid, pre-defined categorization rules that would require manual configuration for each entity, the ML model dynamically adapts to different identifier systems by learning patterns from aggregated historical data. This dynamic categorization reduces the manual effort required for data standardization while enabling efficient recommendation generation.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS11934967B2Providing component recommendation using machine learning

Publication Date: 2024.03.19 TEKION CORP

AI summary

A management system operates in conjunction with entities to provide component recommendations for objects. The management system trains a machine learning model used to generate the component recommendations. The machine learning model is trained based on historical component entries describing components previously provided and identifiers of the components. The management system generates training data by classifying the historical component entries into predetermined component classifications. After the machine learning model is trained, the management system generates a customized recommendation of components for an object based on likelihoods of selection of the predetermined component classifications.