Entity Profile Filtering Using Multi-Model Influence Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for identifying influential individuals in online communities are inefficient, time-consuming, and prone to inaccuracies, often missing influential individuals or falsely identifying them due to reliance on limited data subsets.
Innovation Solution
A multi-model machine learning architecture that processes clinical and online interaction data to calculate influence scores for entity profiles, using a network graph data structure and filtering algorithms to identify current and future influential medical entities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional manual desk research methods are used to identify influential individuals, then the process can be performed with simple tools, but the method is time-consuming and inefficient
Solution Approach 1:
The patent replaces manual mechanical research processes with automated machine learning models. Multiple ML models process clinical data, online interaction data, and network graph data structures to automatically calculate influence scores and identify key opinion leaders, eliminating the need for time-consuming manual desk research while significantly improving identification efficiency
Solution Approach 2:
The system enables self-service automated identification of influential individuals through multi-model machine learning architecture. The models independently process various data sources, calculate metrics, generate network graphs, and produce influence scores without human intervention, allowing the system to serve itself in identifying key opinion leaders and rising stars
2Reliability
If manual analysis of limited data subsets is performed, then the process is simpler to execute, but inaccuracies occur and influential individuals are missed or falsely identified
Solution Approach 1:
The patent segments the identification process into multiple specialized machine learning models, each handling specific aspects: one model processes clinical data and network graphs, another processes online interaction data, and a third model integrates results to calculate final influence scores. This segmentation allows comprehensive data analysis across multiple dimensions while maintaining manageable model complexity through modular architecture
Solution Approach 2:
The system combines multiple data types (clinical data, online interaction data, network graph data) and multiple model outputs into a composite influence score. This composite approach integrates diverse information sources to create a more reliable and accurate identification system that reduces false positives and negatives compared to single-data-source methods
3Measurement precision
If comprehensive data from multiple sources is analyzed, then identification accuracy improves, but the data processing complexity increases
Solution Approach 1:
The patent divides comprehensive data processing into separate specialized models: one model handles clinical data and network graph analysis, another handles online interaction data, and a final model integrates these results. Each model focuses on specific data types and calculations, improving measurement precision while managing processing complexity through modular segmentation rather than requiring one complex monolithic system
Solution Approach 2:
The patent introduces intermediary components including network graph data structures that mediate between raw clinical data and influence score calculations, and intermediate metric calculations that bridge online interaction data with final scoring. These intermediaries organize and structure comprehensive data from multiple sources, enabling precise influence measurement while reducing overall system complexity through structured data transformation
Data Source
AI summary
Disclosed herein are methods and systems for filtering entity profiles. A method includes receiving, from a plurality of data sources and for a plurality of medical entities, clinical data and online interaction data; generating a plurality of entity profiles for the plurality medical entities; calculating one or more clinical metrics and one or more online interaction metrics for the plurality of entity profiles from the clinical data and the online interaction data; executing a model using identifications of the plurality of entity profiles and the one or more clinical metrics and the one or more online interaction metrics as input to generate influence scores for the plurality of entity profiles; selecting a subset of the plurality of entity profiles responsive to each entity profile of the subset having an influence score satisfying a selection criteria; and generating a record comprising identifications of the subset of the plurality of entity profiles.