A rule-based semantic approach for data integration, standardization and dimensionality reduction utilizing the UMLS:...

Area of Science:

• Health Informatics
• Data Science
• Clinical Research

Background:

• Clinical data integration faces challenges due to data heterogeneity and lack of standardization.
• Analyzing large clinical datasets is computationally intensive and time-consuming.
• Existing methods struggle with semantic inconsistencies across diverse data sources.

Purpose of the Study:

• To present a robust method for semantic data integration, standardization, and dimensionality reduction.
• To address challenges in utilizing clinical data for improving patient outcomes.
• To enable efficient data mining on large clinical datasets.

Main Methods:

• Semantic data integration resolving canonical inconsistencies and semantic heterogeneity using the Unified Medical Language System (UMLS).
• Standardization of medical data from diverse sources.
• Dimensionality reduction through a combination of rule-based semantic networks and machine learning.

Main Results:

• Successfully integrated and standardized clinical data from disparate sources.
• Achieved significant data dimensionality reduction, enabling faster data mining.
• Demonstrated the method's utility in predicting bariatric surgery outcomes.

Conclusions:

• The developed method offers a robust solution for clinical data integration and standardization.
• Dimensionality reduction significantly enhances the efficiency of data mining on large clinical datasets.
• This approach facilitates improved patient outcomes through better utilization of clinical data.