
DecicionTree.jl is a CART (Classification and Regression Tree) package in the Julia machine learning programing language. Various steps like loading data, segregating data into training and testing, and building a model is very easy to implement using this package in Julia. There is no need to transpose data as DecisionTree.jl does not require it. It just takes a two lines statement to build the model providing a few arguments (e.g. depth, minimum number of samples in leaf, minimum number of samples in split etc.) as per user requirement. This package saves lots of time and money for regression and classification problems compared to creating a decision tree ML model from scratch. We used decisionTree.jl for a retail client to build a model to classify potential customer data and were able to achieve near about 70% accuracy, which is considered as very good accuracy. Review collected by and hosted on G2.com.
There is no issue as such with DecisionTree.jl as it works smoothly for simple to medium complex problems with good accuracy. I have yet to test its performance for complex problems where input data is very diverse, with more than 20 attributes to consider for classification or regression. One small disadvantage for programmers is that DecisionTree.jl package does not assign data type automatically to your loaded data. It seems it does not assign a data type to improve performance. So, you need to write a few more statements to caste data type to your data before building a model. Review collected by and hosted on G2.com.
At G2, we prefer fresh reviews and we like to follow up with reviewers. They may not have updated their review text, but have updated their review.
Validated through a business email account
This reviewer was offered a nominal gift card as thank you for completing this review.
Invitation from G2. This reviewer was offered a nominal gift card as thank you for completing this review.

