Klasifikasi Sentimen Masyarakat terhadap Kepolisian Negara Republik Indonesia menggunakan Naive Bayes Classifier dan Support Vector Machine

Bagaskara, Nada Damai (2022) Klasifikasi Sentimen Masyarakat terhadap Kepolisian Negara Republik Indonesia menggunakan Naive Bayes Classifier dan Support Vector Machine. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043211117-Undergraduate_Thesis.pdf Download (2MB)

Abstract

Pada tahun 2021 sejumlah anggota kepolisian dinilai beberapa kali melakukan tindakan diluar kode etik dan norma kemanusiaan. Puncaknya adalah akhir tahun 2021 dengan munculnya tagar #percumalaporpolisi, kemunculan tagar ini didorong karena peristiwa tindakan asusila dari beberapa orang anggota polisi. Berdasarkan kejadian-kejadian tersebut, perlu diketahui sentimen masyarakat terhadap Kepolisian Negara Republik Indonesia. Hal tersebut dilakukan karena lembaga Kepolisian Negara Republik Indonesia haruslah memiliki wibawa dan kepercayaan dimata masyarakat, sehingga perlu diketahui ada tidaknya kepercayaan masyarakat terhadap lembaga tersebut. Kepolisian Negara Republik Indonesia sebagai sebuah lembaga perlu menjaga kepercayaan masyarakat, dimana disisi lain masyarakat juga perlu menyampaikan kritik agar lembaga kepolisian tersebut berbenah. Pada penelitian ini dilakukan analisis sentimen dimana data diambil dengan crawling Twitter, menggunakan kata kunci polisi dan polri untuk mendapatkan ulasan masyarakat mengenai kepolisian. Klasifikasi yang digunakan adalah Naïve Bayes Classifier (NBC) dan Support Vector Machine (SVM), dimana selanjutnya dibandingkan untuk mendapatkan metode klasifikasi terbaik. Hasil dari penelitian ini menunjukan bahwa sentimen negatif jauh lebih banyak muncul dibandingkan positif, dengan persentase 80% banding 20%, dimana NBC lebih unggul daripada SVM dengan skor AUC sebesar 0,6410 dan 0,6361 pada ketepatan klasifikasi data training dan data testing.
==============================================================================================================================
In the year of 2021 a number of police officers were assessed several times for taking actions outside the code of ethics and humanitarian norms. The climax is at the end of 2021 with the emergence of the hashtag #percumalaporpolisi, the emergence of this hashtag was driven by the immoral acts of several police officers. Based on these incidents, it is necessary to know the public sentiment towards the of the Indonesian National Police. This is done because the National Police of the Republic of Indonesia must have authority and trust in the eyes of the public, so it is necessary to know whether there is public trust in the institution. The National Police of the Republic of Indonesia as an institution needs to maintain public trust, where on the other hand the public also needs to express criticism so that the police institution can improve. In this study, a sentiment analysis was carried out where the data was taken by crawling Twitter, using the keywords police and police to get public reviews about the of the police. The classifications used are Naïve Bayes Classifier (NBC) and Support Vector Machine (SVM), which are then compared to obtain the best classification method. The results of this study indicate that negative sentiments are much more prevalent than positive, with a percentage of 80% versus 20%, where NBC is superior to SVM with AUC scores of 0.6410 and 0.6361 on the accuracy of training data classification and testing data.

Item Type:	Thesis (Other)
Additional Information:	RSSB 519.53 Bag k-1 2022
Uncontrolled Keywords:	Analisis Sentimen, Naïve Bayes Classifier, Support Vector Machine, Naïve Bayes Classifier, Sentiment Analysis, Support Vector Machine
Subjects:	Q Science > QA Mathematics > QA141 Numeracy--Problems, exercises, etc.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Mr. Marsudiyana -
Date Deposited:	21 Nov 2024 03:02
Last Modified:	21 Nov 2024 03:02
URI:	http://repository.its.ac.id/id/eprint/115806

Actions (login required)

View Item