3
$\begingroup$

This question already has an answer here:

I am not sure if this is the right forum to ask this.

I have some data of the houses, like their size(in square meters), if they use aircondition, how many residents live in, I have their electricity consumption as well. I want to train any Machine Learning Algorithm to the dataset above, in order to create a model that estimates the houses consumption.

I tried many different algorithms (using weka), but I did not have good results. I was said that SVMs could solve this problem, with the right preprocessing. However, i did not have good results either.

Can anyone help me, in the way i should approach this problem?

Thanks in advance

$\endgroup$

marked as duplicate by Raphael Apr 24 '13 at 6:46

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ try "crossvalidated" stackexchange for statistics and datamining. note all approaches will fail if the data is "noisy" and so there is some need to estimate how random the data is. $\endgroup$ – vzn Mar 10 '13 at 16:38
1
$\begingroup$

There are not an appropiate algorithm for each problem, you must to research and compare results between your algorithms or techniques with an error measure (MSE, etc.).

In general, machine learning algorithms are empiric models (approximated functions). There are arised from the problem itself and a set of suppositions with no revealing the hidden model (many of them, are non-linear problems).

First of all, is necessary to be careful with the variables in the input and output layer, this can be categorical, symbolic or numeric (best case), and you have to model your own representation to preserve some features of interest (topology, etc.)

I will recommend you first of all, try with a multiple linear regression and explore the eventual correlations between variables to find any linear dependence, before do something.

Its a very good practice, normalize the numerical data in the training process, to get a well fitted bias.

Then, for instance, you could try with a multilayer feed forward neural network or SVD o fuzzy inference system.

I understand the machine learning problem as a research with a set of methods to get a good explanation for a given problem, and improvements on that methods.

$\endgroup$
  • $\begingroup$ Thank you very much mr arnaldog. I am asking here, because I have tryied almost everything, and I get very bad results.... Thats the reason I wanted something more specific... $\endgroup$ – user7196 Mar 10 '13 at 0:54
  • $\begingroup$ I'm not sure whether this should be merged with the new edition; you may want to (re)post an updated answer there. $\endgroup$ – Raphael Apr 24 '13 at 6:47