I have a regression problem, with a million rows or so, around 10-15 features. What should work better on that particular setting? Neural network or regular regressors?
-
1$\begingroup$ Try using XGBoost with Pandas and Python. Its more light weight with less complexity. If that doesn't perofrm well then you could go for other networks. $\endgroup$– Sangathamilan RavichandranMay 21, 2019 at 8:36
-
$\begingroup$ I'll give it a try thanks! $\endgroup$– BlenzMay 21, 2019 at 8:51
-
2$\begingroup$ Instead of XGboost, I would recommend lightgbm or catboost. Catboost also can deal with factors quite well. Both are faster and less demaning in terms of resources (RAM etc). It really depends on what you want to do, however, I would first of all start with OLS regression. Should be no problem with your data. Always is fast and gives a proper baseline. $\endgroup$– PeterMay 21, 2019 at 8:59
1 Answer
This is more of question how to select the correct machine learning algorithm, I would refer you to the following blog Which machine learning algorithm should I use?
Regression Algorithms models the relationship between variables that is iteratively refined using a measure of error in the predictions. Most popular examples are:
- Ordinary Least Squares Regression (OLSR)
- Linear Regression
- Logistic Regression
- etc ...
On the other hand, Artificial Neural Networks models are inspired by the structure and/or function of biological neural networks. "Neural networks currently provide the best solutions to many problems in image recognition, speech recognition, and natural language processing." Neural Networks and Deep Learning/. Neural Networks are hard to train; thus my recommentation not to start with Neutral Network.