Options
Help: One class classifiers and training data
Contributor I
Hello guys!
I'm trying to evaluate some data. I have an table with "Attempt"; "Username"; "Pressure"; "Pressure_min"; "Pressure_max"; "Size"; "Size_min"; "Size_max". I have 17 different users. 500 attempt per user.
attempt username pressure pressure_min pressure_max size size_min size_max
1 alice 0,845833361 0,837500035762786 0,850000023841857 0,377777804931005 0,333333343267441 0,400000035762787
2 alice 0,829166691 0,812500000000000 0,837500035762786 0,333333343267441 0,333333343267441 0,333333343267441
1 bob 0,80583334 0,562500000000000 0,887499988079071 0,333333343267441 0,333333343267441 0,333333343267441
...
I have the order to classify this with a one class classifier and tried serval things but I still didn't get it how I can do it.
Someone said to me that I have to train the data. So I should take for about 100 attempts form user alice and train them but how? I want to identify the users and make the best accurancy I can get with an one class classifier. When I'm importing the data I leave out the attempt and set the username as a label because I want to classify the user.
Can someone help me, explain me how I train data and how the process must look like? Which classifiers are one class classifiers?
Thank you in advance.
I'm trying to evaluate some data. I have an table with "Attempt"; "Username"; "Pressure"; "Pressure_min"; "Pressure_max"; "Size"; "Size_min"; "Size_max". I have 17 different users. 500 attempt per user.
attempt username pressure pressure_min pressure_max size size_min size_max
1 alice 0,845833361 0,837500035762786 0,850000023841857 0,377777804931005 0,333333343267441 0,400000035762787
2 alice 0,829166691 0,812500000000000 0,837500035762786 0,333333343267441 0,333333343267441 0,333333343267441
1 bob 0,80583334 0,562500000000000 0,887499988079071 0,333333343267441 0,333333343267441 0,333333343267441
...
I have the order to classify this with a one class classifier and tried serval things but I still didn't get it how I can do it.
Someone said to me that I have to train the data. So I should take for about 100 attempts form user alice and train them but how? I want to identify the users and make the best accurancy I can get with an one class classifier. When I'm importing the data I leave out the attempt and set the username as a label because I want to classify the user.
Can someone help me, explain me how I train data and how the process must look like? Which classifiers are one class classifiers?
Thank you in advance.
Tagged:
0
