ABCModeller

ABCModeller performs binary classification of biological datasets by extracting significant features and building ensemble classifiers to support genomics, proteomics, and systems biology analyses.

Key Features:

• Automated Data Preprocessing: Automates preprocessing tasks to clean and prepare biological datasets for downstream analysis.
• Significant Feature Extraction: Identifies significant features using Fibonacci search and orthogonal experimental design.
• Classification Modeling: Constructs classification models using artificial neural networks (ANN), support vector machines (SVM), and random forests (RF).
• Consistent Voting Method: Aggregates predictions from ANN, SVM, and RF with a consistent voting ensemble to improve generalization.
• Model Evaluation and Prediction: Provides evaluation metrics for model assessment and supports prediction on new datasets.
• Hyperparameter Optimization: Automatically selects optimal hyperparameters for machine-learning algorithms via search methods.

Scientific Applications:

• Genomics: Applies binary classification and feature selection to genomics datasets.
• Proteomics: Supports binary classification and feature identification in proteomics datasets.
• Systems Biology: Facilitates modeling and classification tasks within systems biology studies.

Methodology:

Uses automated data preprocessing; feature extraction via Fibonacci search and orthogonal experimental design; classification with ANN, SVM, and RF combined by a consistent voting method; automated hyperparameter optimization; and evaluation metrics for model assessment and prediction.