Classification prediction - Delivering to the business users


#1

Hi All,

Good day. While i was going through solving the classification problems (using Decision Trees) , I understand that we do plot using fancyRpartPlot or any other techniques. That will give you a decision tree with nodes and conditions to arrive at the result . I would like to know how these decision trees will be delivered to the end client in a real time environment . My dumb understanding is we will present the plot to the business users, if we have millions of data how will the end client seggregate it as Yes or No (lets say its a binary classification) for each record. Kindly help.

Thank you.


#2

Hi Nagu,

Along with the plots, we would be also giving the rules from the decision tree. The rules could be derived by following the path to each of the leaf nodes. So whenever a new data point comes in, it will be passed through these rules and the result of the rule (‘yes’ or ‘no’ in your case) which satisfies the given data point will be assigned. This way, the client will be able to classify the data points which they get in.

Hope this helps.!

Thanks,
SRK