This week I updated the correlation detection algorithm to provide more accurate results after viewing fewer examples. Previously, it was only giving accurate predictions after viewing about 1000 different calendar events. Now, after viewing about 300 events, based on my simulation, it can predict about 25% of future stressful events and 80% of those have stress inducing factors. After 1000 events, it can predict 30% of future stressful events and 95% have stress inducing factors. It continues to improve well beyond 1000 events as well. Notable files include Node.java, CorrelationGraph.java, and EventGraph.java.
@poorvi-vora
This week I updated the correlation detection algorithm to provide more accurate results after viewing fewer examples. Previously, it was only giving accurate predictions after viewing about 1000 different calendar events. Now, after viewing about 300 events, based on my simulation, it can predict about 25% of future stressful events and 80% of those have stress inducing factors. After 1000 events, it can predict 30% of future stressful events and 95% have stress inducing factors. It continues to improve well beyond 1000 events as well. Notable files include Node.java, CorrelationGraph.java, and EventGraph.java.