Data Structure
Networking
RDBMS
Operating System
Java
MS Excel
iOS
HTML
CSS
Android
Python
C Programming
C++
C#
MongoDB
MySQL
Javascript
PHP
Physics
Chemistry
Biology
Mathematics
English
Economics
Psychology
Social Studies
Fashion Studies
Legal Studies
- Selected Reading
- UPSC IAS Exams Notes
- Developer's Best Practices
- Questions and Answers
- Effective Resume Writing
- HR Interview Questions
- Computer Glossary
- Who is Who
What are Single-Attribute Evaluators in data mining?
In single-attribute evaluators, it can be utilized with the Ranker search methods to make a ranked list from which ranker discards a given number. It is also used in the RankSearch method.
Relief Attribute Eval is instance-based − It samples instances randomly and checks neighboring instances of the equal and multiple classes. It works on discrete and continuous class data. Parameters define the multiple instances to sample, the various neighbors to check, whether to weight neighbors by distance, and an exponential function that conducts how increasingly weights decay with distance.
InfoGain Attribute Eval − It computes attributes by calculating their information gain concerning the class. It discretizes numeric attributes first utilizing the MDL-based discretization approaches. This approach, along with the following three, can consider missing as an independent value or distribute the counts between other values in proportion to their frequency.
Chi-Squared Attribute Eval − It computes attributes by calculating the chi-squared statistic concerning the class.
Gain-Ratio Attribute Eval − It computes attributes by computing their gain ratio concerning the class.
Symmetrical Uncert-Attribute Eval − It computes an attribute by calculating its symmetrical uncertainty concerning the class.
OneRAttribute Eval − It needs the clean accuracy measure adopted by the OneR classifier. It can need the training data for computation, as OneR does, or it can use internal cross-validation − The multiple folds is a parameter. It can select OneR’s simple discretization approach − The minimal bucket size is a parameter.
SVM-Attribute Eval − It computes attributes using recursive feature removal with a linear support vector machine. Attributes are chosen one by one depending on the size of their coefficients, relearning after everyone.
Indeed, a proportion can be used until a certain number of attributes remain, thereupon switching to the fixed-number method rapidly eliminating many attributes and then considering each remaining one more intensively.
Various parameters are passed on to the support vector machine − complexity, epsilon, tolerance, and the filtering method used.
Principal Components and Latent Semantic Analysis transform the set of attributes. In the case of Principal Components, the new attributes are ranked in order of their eigenvalues. Optionally, a subset is selected by choosing sufficient eigenvectors to account for a given proportion of the variance (95% by default). Finally, the reduced data can be transformed back to the original space.
Latent Semantic Analysis applies a singular value decomposition to the training data. Singular value decomposition is related to principal components analysis-both produce directions that are linear combinations of the original attribute values but differ in that it is computed from a matrix containing the original data values rather than the attribute correlation or covariance matrix.
108 Views
