In feature selection, if the target value is normalized (to number between one and zero) it gives the error value " Unknown label type: 'continuous' ". But if this target value is number other than the decimal between zero & 1 the program can work.
Related to your query, I suggest https://machinelearningmastery.com/how-to-fix-futurewarning-messages-in-scikit-learn/
Muhammad Ali Thank you for your cooperation, sorry it is not warning, it is " ValueError: Unknown label type: 'continuous'".
Your question is not very clear. For example, are you trying to solve a classification problem or regression? Based on what I can get from your question, is that you are trying to solve a multi class (more than two classes) classification problem. In that case, the target values represents the class to which a particular sample is associated with. Hence, you should not be normalizing the target values. Also, there is no need to normalize the target value.
An example would be, for example, if we consider the feature selection class of sklearn ( https://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.SelectKBest.html ), you will see several feature selection methods like f_classif, mutual_info_regression etc. Now f_classif is applicable for classification problem. If your target value is continuous value instead of discrete, you will encounter this error. In cases where your target values are continuous,, mutuak_info_regression is a better choice.
Rajdeep Kumar Nath Thank you, it is not classification problem. It is continuous data.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python