COVID-Predictor

COVID-Predictor predicts coronavirus types from RNA sequence data using a Naïve Bayes classifier to classify SARS-CoV-1, MERS, SARS-CoV-2 and other viruses.

Key Features:

• Machine Learning Approach: Employs a Naïve Bayes classifier to predict coronavirus types from RNA-derived features.
• Training Data: Trained on 1000 RNA sequences including SARS-CoV-1, MERS, SARS-CoV-2 and other viruses.
• Feature Vector Construction: Constructs feature vectors from motifs using k-mer and n-gram techniques.
• Validation and Performance: Validated with 10-fold cross-validation and reported average accuracy of 97% on unseen validation sets.

Scientific Applications:

• Virus Classification: Classifies known and divergent coronavirus sequences such as SARS-CoV-1, MERS, and SARS-CoV-2.
• Viral Evolution Analysis: Supports analysis of genetic patterns and motifs relevant to coronavirus evolution.
• Surveillance and Early Detection: Aids early detection and response strategies by predicting virus types from sequence data.

Methodology:

Data collection of RNA sequences; feature extraction via k-mer and n-gram motif generation to form feature vectors; model training using a Naïve Bayes classifier on the extracted features; validation by 10-fold cross-validation with reported average accuracy of 97%.