I need to build a model using Weka to forecast the future trends of a system
The data feeds I am getting is a vector that comes in pairs. Example
1001,1002,1160,101,111,29-12-2016
1011,1003,1180,102,122,21-11-2016
1101,1004,1110,103,123,27-06-2016
1111,1001,1110,104,111,20-12-2016
1011,1009,1110,105,114,15-09-2016
1101,1102,1150,102,113,01-03-2016
1111,1302,1120,107,118,08-05-2016
1111,1072,1130,101,122,12-10-2016
I have up to a million data set coming in from the unit on monthly basis.
I need to build a cluster to find partern on the data set based on the day of the month.
Secondly I need to try and predict the what the next daily average of the vector could be based on test data.
Example given 2 sample data set. If I validate my sample with the model. I need to get the result as a confidence level or the probability of the vector belonging to a particular cluster or being the next result set.
1001,1002,1160,101,110,29-12-2016 --> will give a confidece of 0.91
1111,1000,1111,102,122,21-11-2016 --> will give a confidece of 0.652
I need to use python and Wake library to run the request. If possible. Could any other system like R do the Job?
