Adding to Luis answer: OpenML currently contains many results at least for Weka. 

For R we have 99% percent of the structure in place to do the same. This would then allow a comparison of many ML algorithms, across tools. The algorithms are evaluated with a wide range of performance measures, we are also interested in timing comparisons. 

We are optimistic that the result database will grow a lot this year and we can also offer summarizing reports / analysis.

The following publications might help you.

https://www.researchgate.net/profile/Mohammed_Al-Kabi/publication/251422102_A_Comparison_Study_between_Data_Mining_Tools_over_some_Classification_Methods/links/00b495345aa85dd423000000.pdf

http://dl.acm.org/citation.cfm?id=2701152

http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=7162956&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D7162956

http://ermt.net/docs/papers/Volume_2/Issue_2_February2013/V2N2-118.pdf

http://onlinelibrary.wiley.com/doi/10.1002/widm.24/full

All the best.

Article A Comparison Study between Data Mining Tools over some Class...

You might also want to look into the no-free-lunch theorems. You cannot have a universal machine learning algorithm that will perform well on every data set, so any comparison can only be restricted on some domain-specific application.

please refer the following links

https://www.google.co.in/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&cad=rja&uact=8&sqi=2&ved=0CDcQFjADahUKEwiviNaIyb7IAhXOno4KHYpVAfI&url=http%3A%2F%2Fwebpages.uncc.edu%2Fmirsad%2Fitcs6265%2FtoolComparison.rtf&usg=AFQjCNGWOPoGRRKDYml-d0Iycd45_3dEBw&bvm=bv.104819420,d.c2E

www.melekirmak.com/trepon/images/260220121521451.pdf

Thanks Taher Ahmed Ghaleb for providing good links...

Thanks Peter Wittek, I got your point of domain-specific application.

thanks Athiraja Atheeswaran

Hello Dr. Parsania, here is the URL of the paper that matches closely to your question:

http://www.cs.uvm.edu/~icdm/algorithms/10Algorithms-08.pdf

If you know the insight of an algorithm then easy to pick a tool for your research.

Thanks.

Hi Vaishali,

as Peter pointed above, the ability or performance of any ML model depends on the specific problem you are working.

FYI, I include a link with one of my works in which six different ML based predictive models are compared for forecasting hourly water demand.

Cheers!

Article Predictive models for forecasting hourly urban water demand

Hi Vaishali,

I know you asked for a research paper, but it may be interesting for you to explore OpenML (http://www.openml.org/) that is a collaborative and open effort to build a huge repository of experiments with different algorithms in different tasks. It allows you to obtain interesting comparative results on a very large set of algorithms on a equally large set of tasks, and moreover you may freely download and re-use the workflows. More on this project in the paper at the link I include.

Cheers

Article OpenML: Networked science in machine learning

Adding to Luis answer: OpenML currently contains many results at least for Weka. 

For R we have 99% percent of the structure in place to do the same. This would then allow a comparison of many ML algorithms, across tools. The algorithms are evaluated with a wide range of performance measures, we are also interested in timing comparisons. 

We are optimistic that the result database will grow a lot this year and we can also offer summarizing reports / analysis.

Adding to Bernd's and Luis' answer: OpenML also has many results for MOA. RapidMiner and Python (scikitLearn) are expected later this year.

If anybody is interested in helping, let us know ;).

Hi Luis and Bernd,

OpenML (http://www.openml.org/) is a very interesting proposal. I will try it!

> if anybody is interested in helping, let us know ;).

+1 Please do so :)

Look at this paper :

http://www.ijarcsse.com/docs/papers/Volume_4/6_June2014/V4I6-0145.pdf

We have tried to touch upon this from the perspective of feature selection It can be read @ Feature Selection: A Practitioner View http://www.mecs-press.org/ijitcs/ijitcs-v6-n11/v6n11-10.html

Outside the research domain there is a study done by Dr, Karl Rexer , which I found to be useful, http://www.rexeranalytics.com/.   You may take a look.

There are so many papers on this area. In fact I am also working on this topic but specifically on a domain rather than general tools

Thanks Saptarsi Goswami for providing link...

Thanks Prof. José Alejandro Lugo García  for providing useful link....

Thanks Prof. Osman Abbas...links of your research paper is useful...

Thanks Laurent Borderie  fro useful link....