G2

Meliksah T.

Engineer

Enterprise (> 1000 emp.)

9/20/2019

"An Extended Machine Learning Tool which contains tools others don't"

5/5

What do you like best about Mlxtend?

I loved its frequent patterns tools apriori and association rules because other common libraries did not have it back then and when I could find those in Mlxtend which was easy to implement, I was so happy.

I also liked how easy it was create ensembled models with Mlxtend's VoteClassifier tools where I was able to test both soft and hard voting for my classification problems. Review collected by and hosted on G2.com.

What do you dislike about Mlxtend?

Even though it does not take huge preprocessing effort before using apriori and association rules functions, it does require some. Besides the format was not explicitly given in the documentation so I spent time on this. Review collected by and hosted on G2.com.

Recommendations to others considering Mlxtend:

VoteClassifer is a good tool but if your data is big, then re-training every model will take time so consider "Dynamic Programming", saving the learned result follow a more manual approach. Review collected by and hosted on G2.com.

What problems is Mlxtend solving and how is that benefiting you?

I used Mlxtend for its frequen patterns tool in the first place, using apriori and association rules algorithms where I looked for the frequent purchases of customers. It was simple and fun to use since it did not require that much in terms of formatting and preprocessing.

Then I used Mlxtend during my machine learning projects to ensemble multiple models. For instance, it has EnsembleVoteClassifier which can do both "hard" and "soft" voting during classification problems. Review collected by and hosted on G2.com.