We are preparing a study on the comparison of regression analysis and decision trees. We compared linear and logistic regression with classification and regression trees on the same data set. We are not just focusing on the statistical point of view but also the easiness and being interpretable for practitioners. What do you think about this? Can you suggest some articles to read on this subject?
In Matlab you could use the "Regression Learner"-App. With this app you can even use machine learning approaches like svm and compare all regression methods.
Hello Ipek,
Decision trees would definitely provide more elaborate and better results than linear or logistic regression. This is because it can partition the data in any number of classes, provided you have sufficient data to do training and testing. From the understanding perspective, Linear/ logistic regression would be the easier one, although decision trees is very easy to understand as well. For the methods mentioned by you, the following article will help you to understand and come to a decision:
https://blog.bigml.com/2016/09/28/logistic-regression-versus-decision-trees/
If you have sufficient data and want to surpass the performance of even the decision trees, then you should probably also look into the random forest method. Probably you may want to try other classification/ regression techniques as well. For that, an easy way is to open the classification or regression learner app in matlab and follow the tutorials. It will be easy to use and compare performance with several techniques (as also suggested by Jann). This toolbox has good inbuilt visuals to demonstrate the performance of various classifiers/ regressors.
Refer to following links if you are using matlab:
https://in.mathworks.com/help/stats/train-classification-models-in-classification-learner-app.html
https://in.mathworks.com/help/stats/train-regression-models-in-regression-learner-app.html
Good luck!
Hello Lpek,
Both of them as an unified tool! Knowing that a strong interconnection links them together...
Follow:
- Criminisi et al., " Decision Forests for Classification, Regression, Density Estimation, Manifold Learning and Semi-Supervised Learning ", Technical Report, 2011 - https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/decisionForests_MSR_TR_2011_114.pdf
- G. Biau et al., " Neural Random Forests ", 2016 - https://hal.archives-ouvertes.fr/hal-01306340/document
- G. Biau et al., " A Random Forest Guided Tour ", 2016 - http://www.lsta.upmc.fr/BIAU/test-bs.pdf
Regards
From an applicative viewpoint, regression analysis is preferred for large scale data analysis and decision trees are well-founded for analyzing small datasets.
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification