Interface administrators, Administrators (Semantic MediaWiki), Curators (Semantic MediaWiki), Editors (Semantic MediaWiki), Suppressors, Administrators
7,785
edits
Feature cross: Difference between revisions
no edit summary
No edit summary |
No edit summary |
||
| Line 1: | Line 1: | ||
{{see also|Machine learning terms}} | |||
==Introduction== | ==Introduction== | ||
Machine learning | [[Machine learning]] [[algorithm]]s rely heavily on [[features]] to extract useful information from [[data]]. [[Feature engineering]] is the process of selecting and manipulating raw data to create new ones that better depict patterns and relationships present in it. One advanced feature engineering technique used is [[feature crossing]]. | ||
==What is Feature Crossing?== | ==What is Feature Crossing?== | ||
Feature crossing is a technique for creating new features by combining two or more existing ones in a dataset. In other words, it captures the interaction between these features. This approach works especially well when the relationship between them is nonlinear or individual elements are not enough to fully capture patterns present in data. | Feature crossing is a technique for creating new features by combining two or more existing ones in a [[dataset]]. In other words, it captures the interaction between these features. This approach works especially well when the relationship between them is nonlinear or individual elements are not enough to fully capture patterns present in data. | ||
Feature crossing is the process of creating new features by multiplying or adding two or more existing ones. For instance, if we have two features A and B, then we can create C by multiplying A by B; that is, C = A*B. Alternatively, adding A to B results in D as D = A + B. | Feature crossing is the process of creating new features by multiplying or adding two or more existing ones. For instance, if we have two features A and B, then we can create C by multiplying A by B; that is, C = A*B. Alternatively, adding A to B results in D as D = A + B. | ||
| Line 26: | Line 27: | ||
==Explain Like I'm 5 (ELI5)== | ==Explain Like I'm 5 (ELI5)== | ||
Machine learning uses feature crossing to create new ways of looking at old data in order to help computers find answers to questions. It's like mixing together your favorite foods to create something even better than the originals! By combining two or more pieces of information and applying some math, we can generate new data which is more helpful in answering our inquiries. | Machine learning uses feature crossing to create new ways of looking at old data in order to help computers find answers to questions. It's like mixing together your favorite foods to create something even better than the originals! By combining two or more pieces of information and applying some math, we can generate new data which is more helpful in answering our inquiries. | ||
[[Category:Terms]] [[Category:Machine learning terms]] | |||