Sakshi Babbar of Amity Online at Machine Learning Developers Summit 2019
summary
A small causal model uses a Bayesian network (built with Netica/bnlearn-style methods) linking environmental variables (temperature, rainfall, humidity), regional/date nodes, symptom nodes and an alert decision node; synthetic data for eight regions (training on 2016–17, testing on 2018 three-month windows) is generated from that model, then high-support high-confidence rules are mined and chi-square tested. The pipeline raises alarms within 48 hours on simulated chikungunya outbreaks, yields near-zero false positives and ~97% correct detection on the test set.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 2:12Problem
- 6:00Framework
- 10:02Objectives
- 10:57Methodology
- 12:09Bayesian Networks
- 16:55Why Bayesian Network
- 18:57Supported Bayesian Networks
- 20:23Methodology for Predictingchikungunya
- 22:31Cause Effect Relationships
- 23:24Probabilities
- 24:11Data Set
- 25:18Bayesian Network
- 25:37Current Day
- 26:19Methodology for MLDS
- 27:30Example
- 28:54Rule
- 31:33Data State
- 32:35Epidemic
- 35:09Data Summary
- 35:57Accuracy
- 37:12Improvements
- 38:05Questions
related talks
41mBayesian Network Modeling using R and Python
26mMachine Learning for Personalised Health
06mCityGuard: Citywide Fire Risk Forecasting Using A Machine Learning Approach
23mApplying AI and ML in the Current Pandemic World
49mModeling our future: Advancing climate research with AI
58mImproving disaster response with machine learning
24mWhat's the Buzz with Machine Learning
18mTransforming Financial Decision Intelligence Through Advanced NNs