02-15-2016 06:36 AM
What is differnce between Logistic Regression using SGD and Random Forest implementation in Mahout? Does both algorithim support mapreduce execution?
08-30-2016 09:46 AM
The difference betweem Logistic Regression using SGD and the Random Forest implimentation is that the Random Forest method is an ensembles method that builds a number of classifiers and then averages out, or selects by voting, the best classifier.
With Logistic Regression using SGD only one classifier is computed.
Both methods are used for classification in Mahout and both can handle categorical and continuous features. Both split the input dataset into train and test partitions for evaluation.
The Random Forest implimentation has only a mapreduce execution mode, whilst Logistic regression using SGD has only a single machine execution mode.