Data reduction and data visualization for automatic diagnosis using gene expression and clinical data.

Area of Science:

• Computational biology
• Medical informatics
• Machine learning in healthcare

Background:

• Accurate disease diagnosis typically requires extensive patient history review by domain experts, which is time-consuming and costly.
• Current disease monitoring tools have limitations, necessitating manual expert analysis for precise classification.

Purpose of the Study:

• To develop an automated diagnostic framework utilizing high-dimensional gene expression and clinical data.
• To address the challenges of analyzing and processing computationally expensive high-dimensional data for disease diagnosis.

Main Methods:

• Implementation of a data reduction technique to transform high-dimensional data into a lower-dimensional, meaningful space.
• Application of data visualization methods to embed complex information into 2-D images.
• Utilization of deep learning approaches on these 2-D images for automated diagnosis.

Main Results:

• The proposed framework demonstrates effective performance in automated disease diagnosis.
• Achieved a high prediction Recall value, ranging between 91% and 99% in experimental analyses.
• Successfully reduced data dimensionality while preserving essential information for diagnostic purposes.

Conclusions:

• The developed framework offers an efficient and accurate automated approach to disease diagnosis.
• Deep learning applied to visualized, reduced-dimension data shows promise for clinical applications.
• This method has the potential to significantly reduce the effort and cost associated with disease classification.