can it be used for usual classification problem?

[Sandy4321]

can it be used for usual classification problem when wen each data observation is one vector and when there are 2 classes : yes and no in short can it be used for usual logistic regression cases? like https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html

[tiepvupsu]

Yes, absolutely. I'll make one example for that task.

[Sandy4321]

Great thanks, looking forward so it will be example for 2 classes ( let's say yes and no) for classification were each data observation is presented by vector. Data is unbalanced data let's say 10% label yes and 90% label no Better if this vector is spars and each element is 1 or 0 For example 80% of zeros Then we can see clear pattern to find

For example we may have 100000 observations When yes observations are minority for example 3000 yes observations Then vector length is 150 We may have 15 patterns Each pattern is 20 values long For example pattern 1 1111111111111111111 Pattern 2 0101010101010101010 Etc Then let's run logistics regression from scikit learn and your code Hopefully your code will be better?

[Sandy4321]

I see you updated to Python 3 Great thanks It would be very kind of you to add this well deserved classification example As well Thanks a lot in advance