RETRACTED: Diagnosis labeling with disease-specific characteristics mining

Area of Science:

• Machine Learning
• Data Science
• Computer Science

Background:

• The study aimed to evaluate machine learning models.
• Concerns arose regarding the methodology and data handling.
• Retraction was initiated due to identified errors.

Purpose of the Study:

• To present findings on machine learning model performance.
• To address issues of overfitting and inadequate comparisons.
• To ensure scientific integrity and accurate reporting.

Main Methods:

• Dataset partitioning into training, validation, and testing sets.
• Implementation of algorithms for baseline comparisons.
• Utilizing a bag-of-words model for text representation.

Main Results:

• Overfitting was identified due to parameter tuning on training and validation sets.
• Baseline comparisons were deemed insufficient for robust evaluation.
• The bag-of-words model exhibited signs of overfitting, impacting trustworthiness.

Conclusions:

• The presented results are unreliable due to methodological flaws.
• Overfitting in model training and evaluation compromises findings.
• The study's conclusions are invalidated by data and comparison issues.