Best algo for LOG files multiclass classification?

I have a big dataset containing logs/steps that the user performed on my webpage (for example: Clicking on a “Homepage” button, typing some text in the field, etc.) These steps are labelled against their respective action/class name (each set of steps belong to a specific class), what kind of ML algo can I apply on this kind of problem?

Special NOTE: each of these “classes”/“activities” have multiple steps under them, and I don’t know whether a simple classification algo such as LR or KNN would work…

example data: This is just one class data, and the data is arranged in this format:

{ CLass : [Title>>Header>>Action>>Input_Field , Title2>>Header2>>Action>>Input_Field]}

“Create Change Order from Purchase Order”: [“nan>>nan>>component_name>>nan”, “nan>>nan>>Login To Cloud_Central>>nan”, “nan>>nan>>Login and Get Credentials form Excel_Central>>nan”, “nan>>nan>>Application_TypeByText>>User ID”, “nan>>nan>>Application_TypeByText>>Password”, “nan>>nan>>Application_Click>>Sign In”, “nan>>nan>>Reset Global Variable>>nan”, “nan>>nan>>Close All Browsers Central>>nan”, “nan>>nan>>Close PDF And Excel Central>>nan”],

I tried a DTW and pattern matching algo for matching an input sequence to the given json data.

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.