I am currently working on highly non-linear data that is very limited to approx. 1000 (integer type data), As I know Random Forest algorithm works very well for data at least 3000~5000. I need any other machine learning algorithm to deal with limited data sets for good prediction.
http://pressurevesseltech.asmedigitalcollection.asme.org/article.aspx?articleid=2615421
Article Random forest in remote sensing: A review of applications an...
Dear Muhammad Abdul Qyyum ,
Look the link, maybe useful.
https://blog.frankfurt-school.de/wp-content/uploads/2018/10/Neural-Networks-vs-Random-Forests.pdf
Regards,
Shafagat
I recommend you to evaluate what is goodness between different model. what is the problem domain, sensitivity issues, unbalancing data challenge, Hughes phenomenon issue, stability and generalization capability of the each model determine whether it fit for the given problem or not.
To classify land use land cover, i utilized XGboost with stacking ensemble learning method and obtained optimal result
It is not easy to say which algorithm better without knowing the exact nature of the data, You should try ANN, k-NN, SVM, XGboost, ELM and Deep learning models.
Classification depends on the data/features given to a specific classifier as an input. I think its better to try classifiers by hit and trial manner to choose the one give you best results. Neural Networks are better choice to test
I think there is no suitable classifier in all cases, this depends on many factors like the type and size of data
you should check up which algorithms are suitable for your data, it is difficult to say which algorithm suitable for your data before you try.
in general, an ensemble learning-based method doing well, you can try it with XGBoost, but you should tuning hyper-parameters to get good performance.
Your did not mentioned the number of inputs metrics and labels. And what will you do? Classification? prediction? Optimization. Etc.?
- Badges
- Science topic