Fire Incident risk classification Data Mining project
Property loss is one of the end effect of a fire incident which can be controlled by taking preventive actions or it can be minimized by responding to the incident in a faster and efficient way by using predictive analysis based on the available data with variables such as incident type, incident time and date, estimated property losses, area zip code and neighborhood. Applying these variables to the data mining concept we can strive towards reduction in response time, property loss and other resource based on prioritization.
Boston fire department website https://data.boston.gov/dataset/fire-incident-reporting
The Logistic regression model employed to classify the new dataset has an accuracy of 0.775