[User Story] A novice user want to know how can his task benifit from the the provided AI solution (DataFactory)

[HuangYiran]

• For each art of task, introduce different evaluation methods and provide blank method for our customer to customize his evaluation method.
• For each art of task, introduce different baseline and provide blank method for our customer to customize his own baseline.
• show the customer the information of his data, including the flaws of the data (na, outliear), the statistic summary (mean, max, min...), distribution of each feature, relationship between features, relationship between the given features and target and so on.
• show the customer the way to fix the flaws and sinple feature engineering according to the attribution discovered, e.g., labeling, clustering, normalizing.
• show the customer the performance of the basic AI method, e.g., linear model, decision tree, svm.
• show the feature importance according to the basic model

[HuangYiran]

@nischa564 for the finetune function, can I use cpu only? Do we have a function, which do not do the finetuning and just compare the result of several basic model e.g., linear, dt, rf with different metrics?

[nischa564]

@nischa564 for the finetune function, can I use cpu only? Do we have a function, which do not do the finetuning and just compare the result of several basic model e.g., linear, dt, rf with different metrics?

@HuangYiran Until now I used gpu as standard device, if one exists. But it is easy to add that, if you want me to.
No, we have no such function until know.

[HuangYiran]

I have added the function to the package (put it at the ts/preprocessing so far
). It will be good that we can run the finetune method (maybe limited version) with cpu.