Text

REDUKSI DIMENSI FITUR DENGAN METODE FEATURE SELECTION INFORMATION GAIN PADA KLASIFIKASI MODIFIED K NEAREST NEIGHBOR

---

Penilaian

---

The Classification algorithm Modified k-Nearest Neighbor (MkNN) is a development of the kNN algorithm where MkNN can solve the outlier problem in ordinary KNN. MkNN has several disadvantages such as requiring large computation and memory costs in its application and not good to dealing high-dimensional data. From this arises the question, whether dimension reduction by feature selection using information gain has an effect that can overcome these weaknesses. Moreover, it is known that feature selection has a direct effect with reduced processing time for data mining algorithms, improves performance in classification and also results that are easier to understand. To determine the effect of this dimension reduction, the method will be tested on dataset LSVT Voice Rehabilitation. MkNN classification using dimensional reduction with Information Gain results in an average accuracy of 83.46%, the average time of 6 seconds and the average memory of 120130765 bytes while MkNN classification without dimensional reduction results in an average accuracy of 80.78%, the average time is 12.5 seconds and the average memory is 121313689 bytes.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2007000924	T39239	T392392020	Central Library (REFERENCES)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T392392020

Publisher

Indralaya : Jurusan Teknik Informatika Fakultas Ilmu Komputer Universtas Sriwijaya., 2020

Collation

xvi,30 hlm.; ilus., tab.: 28 cm

Language

Indonesia

ISBN/ISSN

-

Classification

005.07

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Dimensi Fitur

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment