Predective Analysis

Gurus,

I have come up with the below requirement. The objective is as below. The reason why I am putting this example is beacuse I want to understand how the Predective analysis works. I want to understand how such scenarios can be handled? How to decide which nodes to use and why? What all things we need to consider while performing such predictions.

My example takes in the expenses I made every month till May 2014. I have my expenses figures for rest of the months which have gone by and the future months like Oct, Nov, Dec, Jan - 2015,Feb 2015 I want to predict. I want this time frame to be dynamic like if I want to predict till Feb 2015 at some point of time and later want to predict my expenses till Jun 2015 then I must be allowed to do so with minimum change. My expenses are also dependent on various factors, like shopping, movies, party with friends, etc. I can tell the impact of these in percentage or figure while predicting my expenses for future. Like I am expecting that 5% of my expense will be on shopping for June 2014.

Year Month Sum Year_Month
2013 4 125 2013_4
2013 5 232 2013_5
2013 6 154 2013_6
2013 7 476 2013_7
2013 8 243 2013_8
2013 9 373 2013_9
2013 10 124 2013_10
2013 11 278 2013_11
2013 12 299 2013_12
2014 1 346 2014_1
2014 2 474 2014_2
2014 3 168 2014_3
2014 4 255 2014_4
2014 5 378 2014_5
? ? ? ?
? ? ? ?
? ? ? ?
? ? ? ?
? ? ? ?
? ? ? ?
? ? ? ?

Cheers!!

 

Thanks in advance :)

Hi to use the workflow in attachment, you need the Dumb joiner component from Lhasa limited extension. I did it for a training to show how to use wrapped nodes. So this is a "basic" forecast node.

To use it : right clic on the Forecast node to use parameters (for example number of months to forecast).

Description is in the Node description on the left panel.

Hope it will give you some ideas.

 

Fabien