KLASIFIKASI KOMENTAR SPAM PADA INSTAGRAM MENGGUNAKAN ALGORITMA MUTUAL INFORMATION DAN MODIFIED K-NEAREST NEIGHBOR

INDRIANI, DESI and Abdiansah, Abdiansah (2021) KLASIFIKASI KOMENTAR SPAM PADA INSTAGRAM MENGGUNAKAN ALGORITMA MUTUAL INFORMATION DAN MODIFIED K-NEAREST NEIGHBOR. Undergraduate thesis, Sriwijaya University.

[thumbnail of RAMA_55201_09021181621023.pdf] Text
RAMA_55201_09021181621023.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (5MB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_0001108401_01_front_ref.pdf]
Preview
Text
RAMA_55201_09021181621023_0001108401_01_front_ref.pdf - Accepted Version
Available under License Creative Commons Public Domain Dedication.

Download (3MB) | Preview
[thumbnail of RAMA_55201_09021181621023_0001108401_02.pdf] Text
RAMA_55201_09021181621023_0001108401_02.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (637kB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_0001108401_03.pdf] Text
RAMA_55201_09021181621023_0001108401_03.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (297kB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_0001108401_04.pdf] Text
RAMA_55201_09021181621023_0001108401_04.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (1MB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_0001108401_05.pdf] Text
RAMA_55201_09021181621023_0001108401_05.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (412kB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_0001108401_06.pdf] Text
RAMA_55201_09021181621023_0001108401_06.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (10kB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_0001108401_07_ref.pdf] Text
RAMA_55201_09021181621023_0001108401_07_ref.pdf - Bibliography
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (228kB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_0001108401_08_lamp.pdf] Text
RAMA_55201_09021181621023_0001108401_08_lamp.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (146kB) | Request a copy
[thumbnail of RAMA_55201_09021181621023_TURNITIN.pdf] Text
RAMA_55201_09021181621023_TURNITIN.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (10MB) | Request a copy

Abstract

Spam on Instagram is defined as comments that are unrelated to the photo or video that is being commented on. Spam comments disrupt the flow of the conversation, making it difficult for users to find information quickly and precisely. Some ways that can be done to solve the problem of spam comments are to classify comments based on the category of spam comments or not spam. One method that can be used for classification is Modified K-Nearest Neighbor (MKNN). However, the MKNN method has a weakness in processing high-dimensional data, so a feature selection method is required to reduce the document's number of features. The method used is Mutual Information (MI). The purpose of this study was to determine the performance of the MKNN classification with MI and MKNN without using MI. The results of the test will be compared and evaluated to see the effect of the MI feature selection method in improving MKNN performance. The test is carried out with the input k values of 3, 5, 7, and 9 and the threshold values of 0.018, 0.01, and 0.008. The test results that have been obtained show that the MKNN method with MI has better performance than the MKNN without MI. MI feature selection is able to reduce the number of features in the data set thereby improving classification performance. The difference in the increase in highest accuracy occurs when the value of k = 3 with a threshold value of 0.018, which is 22%.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Instagram, Comments, Modified K-Nearest Neighbor, Mutual Information, Spam
Subjects: T Technology > T Technology (General) > T1-995 Technology (General)
Divisions: 09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User: Desi Indriani
Date Deposited: 14 Sep 2021 08:18
Last Modified: 14 Sep 2021 08:18
URI: http://repository.unsri.ac.id/id/eprint/53970

Actions (login required)

View Item View Item