Aplikasi Data Mining berbasis Web untuk Klasifikasi Impresi pada Caption Pengguna Instagram

Priwi, R. Sidqi Tri (2019) Aplikasi Data Mining berbasis Web untuk Klasifikasi Impresi pada Caption Pengguna Instagram. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05111540000153-Undergraduate_Theses.pdf Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Dalam proses pembuatan profil Instagram pengguna banyak membagikan aktivitas dan perkataan mereka. Melalui caption para pengguna Instagram, banyak kepribadian para pengguna Instragram dapat teridentifikasi melalui profil mereka. Penulis mengusulkan algoritma klasifikasi impresi pada caption pengguna Instagram untuk menampilkan hasil informasi klasifikasi impresi. Metode klasifikasi yang digunakan adalah K-Nearest Neighbour (KNN) dan Support Vector Machine (SVM). Dataset caption pengguna Instagram di dapat dari github.com/SenticNet. Dataset berupa 2467 data caption yang sudah terklasifikasi ke dalam 5 impresi berdasarkan Kepribadian Big Five. Perancangan sistem terdiri dari desain arsitektur sistem, perancangan dataset, perancangan alur algoritma text mining yang terdiri dari Tokenizing, Filtering, Stemming dan Term Weighting TF-IDF, tahap pelatihan dan pengujian dan terakhir perancangan web. Beberapa uji coba telah dilakukan salah satunya menggunakan Cross Validation K-Folds. Nilai akurasi terbaik yang didapat yaitu 55.07% untuk KNN dan 40.58% untuk SVM.
====================================================================================================================================
In the process of creating Instagram profiles many users share their activities and sayings. Through the captions of Instagram users, many of the personality of Instragram users can be identified through their profiles. The author proposes an impression classification algorithm in the Instagram user caption to display the results of impression classification information. The classification method used is K-Nearest Neighbor (KNN) and Support Vector Machine (SVM). The caption dataset for Instagram users is from github.com/SenticNet. The dataset is 2467 caption data that has been classified into 5 impressions based on the Big Five Personality. The system design consists of system architecture design, dataset design, design of text mining algorithms consisting of Tokenizing, Filtering, Stemming and Term Weighting TF-IDF, training and testing stages and finally web design. Several trials have been carried out, one of them using Cross Validation K-Folds. The best accuracy values obtained were 55.07% for KNN and 40.58% for SVM.

Item Type:	Thesis (Other)
Additional Information:	RSIf 006.312 Pri a-1 2019
Uncontrolled Keywords:	Instagram, KNN, SVM, Impresi
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	R. Sidqi Tri Priwi
Date Deposited:	06 Aug 2026 02:09
Last Modified:	06 Aug 2026 02:09
URI:	http://repository.its.ac.id/id/eprint/67791

Actions (login required)

View Item