Machine Learning vs The Flu
summary
Predicts next-season HA and NA sequences from IRDB data, encoding A/T/G/C as numerics in 15-base chunks, and compares decision trees, extra trees, gradient-boosting/XGBoost and random forest. Uses Biopython to read FASTA/GenBank and scikit-learn for cross-validated training and R² scoring. Random forest gives the best ensemble performance versus formula-based studies, but model only uses parent→child generations, underfits fast-mutating H1N1, and hits numeric precision limits; ongoing work integrates codon grouping and more generations.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:24Flu vaccine methods problems
- 3:33Flu infection and proteins
- 5:55Phylogenetic trees approach
- 8:45Biopython discovery
- 10:28Decision trees struggles
- 11:53Scikit-learn modeling
- 13:15Model training process
- 16:30H1N1 and H3N2 results
- 18:38Formula-based comparison
- 19:21Antigenic drift vs shift
- 22:37Future work
- 23:22Q&A
related talks
33mApplication of Multi-class Supervised Algorithms to Identify COVID-19 Viral Strains
42mPython’s role in accelerating biomedical science
42mLearning Machine Learning
48mLearning Machine Learning
39mForecasting With Classical and Machine Learning Methods
25mTimeSeries Forecasting with ML Algorithms and there comparisons
37mAI/ML Strategies & Techniques for a Post Pandemic World
32mMachine Learning-Based Design Of Proteins