Ericsson
Question
Asked 11 March 2016
What is the difference between feature extraction and classification?
where does feature extraction and classification apply?
Most recent answer
As Sachin mentioned, feature selection is part of preprocessing. Be wary of the difference between feature extraction and selection. Feature extraction is the process in which the each given sample is broken down to quantifiable (and sometimes qualitative) attributes. Feature selection is the process of choosing a subset of those extracted attributes for a domain specific application.
Both the above tasks are to be applied on the whole data set. Although technically, the analysis that forms the basis for feature selection is done only on the training set but the resulting outcome is applied to all training, validation and testing sets.
2 Recommendations
Popular answers (1)
Ericsson
Let's be clear what a feature is. Given a set of samples on which you want a make a decision on, a feature is something that would possibly be different among those samples, i.e, discriminate one sample from another. This decision may be to classify the sample to a finite set, group the samples based on similarity (clustering) or predict a real value based on those features (regression).
Feature extraction is the process of collecting discriminative information from a set of samples. Feature classification is the grouping of features based on some criteria. Sometimes feature classification might also be related to feature selection which is to select a subset of the extracted features that would optimise the machine learning algorithm and possible reduce noise removing unrelated features.
To get further information regarding this, you may refer the following paper.
5 Recommendations
All Answers (5)
Ericsson
Let's be clear what a feature is. Given a set of samples on which you want a make a decision on, a feature is something that would possibly be different among those samples, i.e, discriminate one sample from another. This decision may be to classify the sample to a finite set, group the samples based on similarity (clustering) or predict a real value based on those features (regression).
Feature extraction is the process of collecting discriminative information from a set of samples. Feature classification is the grouping of features based on some criteria. Sometimes feature classification might also be related to feature selection which is to select a subset of the extracted features that would optimise the machine learning algorithm and possible reduce noise removing unrelated features.
To get further information regarding this, you may refer the following paper.
5 Recommendations
Amrita Vishwa Vidyapeetham
Thanks for the response Ebenezer R.H.P. Isaac
I would like to ask one more question on this.
Is Feature extraction on different samples of data set or on the entire data set ? If so, what difference does it make?
Feature selection is the data preprocessing task. In feature selection, we try to identify relevant attributes of the data which can be useful for the analysis and removing those attributes which are not so relevant for the analysis.
Classification is a data mining task, which can be applied to the data sets with class labels already known for the target attribute or dependent attribute (after preprocessing). If you don not know the class label information, then go for clustering.
All the Best
2 Recommendations
Ericsson
As Sachin mentioned, feature selection is part of preprocessing. Be wary of the difference between feature extraction and selection. Feature extraction is the process in which the each given sample is broken down to quantifiable (and sometimes qualitative) attributes. Feature selection is the process of choosing a subset of those extracted attributes for a domain specific application.
Both the above tasks are to be applied on the whole data set. Although technically, the analysis that forms the basis for feature selection is done only on the training set but the resulting outcome is applied to all training, validation and testing sets.
2 Recommendations
