Survey on Meta Learning for Choosing Classifier |
Author(s): |
| Prof. Jaiminee Patel , LDRP-ITR, KSV, Gandhinagar, Gujarat; Prof. Kamaxidevi P Raol, LDRP-ITR, KSV, Gandhinagar, Gujarat; Prof. Hardik Patel, LDRP-ITR, KSV, Gandhinagar, Gujarat |
Keywords: |
| Metalearning; Metafeatures; Learning Algorithms |
Abstract |
|
These days, there are a wide range of data classifiers available, making it difficult for those who are not familiar with the properties of data and how they are distributed to know which data classification method should be applied to produce accurate classification results for their specific dataset. Choosing a proper classifier for a given dataset is therefore a crucial challenge. Finding or extracting meta-features-features that define the data itself-from a particular dataset is referred to as meta-learning. In this study, we evaluate five different types of meta features for their usefulness in predicting the classification accuracies of many popular classifiers. |
Other Details |
|
Paper ID: LDRPTCP037 Published in: Conference 12 : LDRP TECON23 Publication Date: 23/12/2023 Page(s): 197-200 |
Article Preview |
|
Download Article |
|
|
