Articles

Found 13 Documents
Search

IDENTIFIKASI KOMENTAR SPAM PADA INSTAGRAM Chrismanto, Antonius Rachmat; Lukito, Yuan
Lontar Komputer Vol. 8, No. 3 Desember 2017
Publisher : Research institutions and Community Service, University of Udayana

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/LKJITI.2017.v08.i03.p08

Abstract

Spam pada Instagram (IG) umumnya berupa komentar yang dianggap mengganggu karena tidak berhubungan dengan foto atau video yang dikomentari. Spam pada komentar dapat menyebabkan beberapa dampak negatif seperti menyulitkan untuk mengikuti diskusi pada komentar yang dipenuhi oleh komentar spam dan menyebabkan seseorang tampak populer karena jumlah komentarnya banyak walaupun pada kenyataannya lebih banyak komentar yang berupa spam. Penelitian ini mencoba untuk membangun model yang dapat melakukan identifikasi komentar spam pada IG. Komentar pada IG berbentuk teks, sehingga pada penelitian ini digunakan metode-metode pengolahan teks.  Untuk identifikasi digunakan metode Support Vector Machine (SVM). Data komentar yang digunakan pada penelitian ini dikumpulkan dari komentar-komentar pada foto atau video yang dibagikan oleh aktor dan artis Indonesia yang memiliki pengikut (follower) paling banyak di IG.  Dari hasil penelitian didapatkan model identifikasi komentar spam dengan metode SVM menghasilkan tingkat akurasi 78,49% yang lebih baik jika dibandingkan dengan model pembanding yang menggunakan metode NB (77,25%). Penelitian ini juga menguji beberapa proporsi data pelatihan yang berbeda-beda dan hasilnya metode SVM tetap lebih baik dibandingkan dengan metode NB. Hasil lain dari penelitian ini adalah tahap pre-processing dan stemming yang harus disesuaikan terutama untuk dukungan terhadap pengolahan karakter-karakter unicode dan simbol-simbol khusus yang banyak ditemukan pada komentar-komentar di IG.
Indoor Positioning System dengan Algoritma K-Means dan KNN Suryanto, Hizkia Juan; C., Antonius Rachmat; Lukito, Yuan
Jurnal Teknik Informatika dan Sistem Informasi Vol 2 No 3 (2016): JuTISI
Publisher : Maranatha University Press

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1553.655 KB) | DOI: 10.28932/jutisi.v2i3.641

Abstract

Indoor Positioning System (IPS) can determine someone’s position inside a building. The common method used is implemented by WiFi signal strength analising. This paper discusses about how to do IPS using K-Means and K-Nearest Neighbor (KNN) method, that also analyze the accuracy. K-Means is used to cluster dataset. Each data in certain cluster then classified using KNN method. The dataset consists of 11658 Received Signal Strength (RSS) from 177 Access Point (AP) in UKDW. Accuracy of system analized using 10-fold Cross Validation method which is applied in a range of k=2 to k=11 for clusterisation process, then k=1 to k=5 for classification process. Based on the experiment results, system can determine someone’s position with 88.49% accuracy which k optimum is 10 for clusterisation process, and k=1 for classification process.
IMPLEMENTASI ALGORITMA NAÏVE BAYES MENGGUNAKAN ISEAR UNTUK KLASIFIKASI EMOSI LIRIK LAGU BERBAHASA INGGRIS Astuti, Laksmita Widya; Rachmat C., Antonius; Lukito, Yuan
Jurnal Informatika Vol 14, No 1 (2017): MAY 2017
Publisher : Institute of Research and Community Outreach - Petra Christian University

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (306.62 KB) | DOI: 10.9744/informatika.14.1.16-21

Abstract

Lirik lagu merupakan suatu ungkapan perasaan seseorang terhadap sesuatu hal yang sudah dilihat, didengar maupun dialaminya sehingga tidak jarang emosi menjadi salah satu kriteria user dalam melakukan pencarian. Pencarian lirik melekat pada kategori yang tidak hanya terbatas berdasarkan genre atau judul lagu, namun juga melalui emosi dari lirik lagu yang diungkapkan. Agar dapat mencapai tujuan tersebut, diperlukan suatu sistem pengkategori yang mengenali lirik lagu secara otomatis dengan salah satu metode klasifikasi yaitu Naïve Bayes. Faktor yang mendorong tingginya tingkat akurasi bukan hanya terletak pada metode klasifikasi saja, namun proses sebelum menuju tahap klasifikasi juga berpengaruh pada hasil yang didapatkan. Maka dari itu, penulis melakukan penelitian melalui beberapa tahap yaitu preprocessing berupa tokenisasi, stopword dan stemming, kemudian feature selection yang digunakan adalah TF-IDF dengan bantuan ISEAR karena mengandung 7 emosi dasar. Dari ketujuh emosi dasar tersebut, tiga diantaranya merupkan emosi yang akan digunakan dalam penelitian ini yaitu anger, sadness dan joy. Hasil dari penelitian ini menunjukkan dengan menggunakan ISEAR akurasi tertinggi terdapat pada feature set 60% dan 100% yaitu sebesar 82,2%. Perbedaan signifikan dihasilkan pada penggunaan ISEAR dengan akurasi rata-rata keseluruhan porsi featureset sebesar 77% sedangkan tanpa menggunakan ISEAR rata-rata akurasi sebesar 53%. Dokumen paling relevan untuk pengujian menggunakan ISEAR terdapat pada kategori angry dengan rata-rata f-measure sebesar 0.7267.
Model Multi Layer Perceptron untuk Indoor Positioning System Berbasis Wi-Fi Lukito, Yuan
Jurnal Teknologi dan Sistem Komputer Volume 5, Issue 3, Year 2017 (July 2017)
Publisher : Departemen Teknik Sistem Komputer, Fakultas Teknik, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (357.523 KB) | DOI: 10.14710/jtsiskom.5.3.2017.123-128

Abstract

Masalah penentuan posisi di dalam ruangan masih memerlukan banyak perbaikan. Penelitian ini mencoba melakukan eksplorasi terhadap penggunaan multi layer perceptron untuk penentuan posisi seseorang di dalam gedung atau ruangan, yang lebih dikenal dengan istilah Indoor Positioning System. Penelitian ini dilaksanakan dalam beberapa tahap yaitu normalisasi dataset, implementasi multi layer perceptron, pelatihan multi layer perceptron dan proses pengujian serta analisis. Proses pelatihan dilakukan beberapa kali untuk menemukan parameter-parameter yang menghasilkan akurasi terbaik. Dari hasil percobaan yang dilakukan didapatkan tingkat akurasi terbaik sebesar 79,16%. Hasil tersebut masih lebih rendah jika dibandingkan dengan hasil penelitian sebelumnya, sehingga memerlukan perubahan pengaturan parameter atau pengubahan arsitektur jaringan syaraf tiruan yang digunakan.
Perbandingan Metode-Metode Klasifikasi untuk Indoor Positioning System Lukito, Yuan; Chrismanto, Antonius R.
Jurnal Teknik Informatika dan Sistem Informasi Vol 1 No 2 (2015): JuTISI
Publisher : Maranatha University Press

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1607.933 KB) | DOI: 10.28932/jutisi.v1i2.579

Abstract

Indoor Positioning System can provide position and navigation guidances inside a building.  This paper discusses about systematic comparison between K-Nearest Neigbors and Naïve Bayes Classifier over WiFi-based Indoor Position System dataset.  The dataset is collected using a custom Android Application, which able to receive and record WiFi signal strengths from the surrounding WiFi hotspots in UKDW campus. The dataset consists of 11658 Received Signal Strength (RSS) data from 41 public locations in UKDW campus.  We use 10-folds cross validation and T-Test with 0.05 significance level to compare classification accuracy between K-Nearest Neigbors and Naïve Bayes classifier.  Based on the experiment result, we can conclude that K-Nearest Neighbors classifier produces better classification accuracy (83.58%) than Naïve Bayes (61.52%).
PENERAPAN SENTIMENT ANALYSIS PADA HASIL EVALUASI DOSEN DENGAN METODE SUPPORT VECTOR MACHINE Santoso, Valonia Inge; Virginia, Gloria; Lukito, Yuan
Jurnal Transformatika Vol 14, No 2 (2017): January 2017
Publisher : Jurusan Teknologi Informasi Universitas Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26623/transformatika.v14i2.439

Abstract

The quality of lectures can be determined by some feedbacks from students. From the feedbacks, we can give appreciations for those lectures who get good feedback from students, and evaluations for those who get bad feedback. The problem is classifying large size of feedbacks manually isn’t effective and took a lot of time. Therefore, we need a system that can classify feedbacks automatically. These feedbacks will be classified into positive, negative, and neutral, usually called as sentiment analysis. Sentiment analysis implementation can be done by several methods, one of them that has a good accuracy is Support Vector Machine (SVM). SVM performance in this research is measured with the level of accuracy. The number of accuracy indicate the success level of system. The conclusion of this research is factors that affects the accuracy. The factors are the range of each classes and number of unique words in the training document.
IMPLEMENTASI ALGORITMA RIJNDAEL 128 PADA APLIKASI CHATTING BERBASIS HTML5 WEBSOCKET Sularso, Eko; Rahardjo, Willy Sudiarto; Lukito, Yuan
Jurnal Informatika Vol 10, No 2 (2014): Jurnal Teknologi Komputer dan Informatika
Publisher : Universitas Kristen Duta Wacana

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1235.532 KB) | DOI: 10.21460/inf.2014.102.326

Abstract

In the past, web-based chat application didn’t consider security as part of must-have requirement, thus many insecure examples were broken in short time after it was released. Data sniffing is one common attack that could be used to attack insecure applications because the data was transferred using an insecure medium, which is HTTP. We propose a new web-based chat application that is built based on HTML5 WebSocket technology using Socket.IO library to improve confidentiality of the messages sent between two or multiple parties. We combine it with NodeJS and Express to facilitate real-time discussion between client and server and vice versa. We also use Rijndael (known as AES - Advanced Encryption Standard) to make sure that the message stays confidential and only known by sender and receiver. To satisfy the integrity property, we apply SHA-3 hash function. By combining SSL/TLS, AES, and SHA-3 hash function, we have added multiple layer of security inside this application and no additional effort needed by the user. Based on conducted experiments, we can conclude that this application could satisfy security requirements (confidentiality and integrity), either on the client or server side.
VERIFIKASI AKUN DATABASE DENGAN PENERAPAN METODE TEMPLATE MATCHING PADA KARATERISTIK WAJAH PERSONAL Pebrindanov, Ginting; Hapsari, Widi; Lukito, Yuan
Jurnal Informatika Vol 11, No 1 (2015): Jurnal Teknologi Komputer dan Informatika
Publisher : Universitas Kristen Duta Wacana

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (4638.858 KB) | DOI: 10.21460/inf.2015.111.427

Abstract

The identification of account is a step to keep the important data secure. Nowadays, it can be done by using username and password, but, after see the reality, the using of username and password can’t keep the data secure from the thief. Because of that, the  verification of the characteristic of personal face can be a solution to change the using username and password. The method that can be used for verification is template matching.It is implemented in four features of personal face, such as left eye, right eye, nose and lips. The four images of  each feature will be extracted with wavelet haar method. The feature extraction will be done during template taking process and verification. The result of this research, the result of verification is determined by two factors,  such as the distance between face and web camera is different when the template taking process and the verification process and the diferrent brightness condition when the template taking process and the verification proccess. The threshold value that has been decided is not really able to block the unregistered data.  Then the accuracy of the verification activity is still low and it is still not able yet to identify an account well.
IMPLEMENTASI DYNAMIC PROGRAMMING PADA PENENTUAN JENIS MATERIAL UTAMA BANGUNAN ARENA FUTSAL Anthony, Andree; Santosa, R. Gunawan; Lukito, Yuan
Jurnal Informatika Vol 10, No 1 (2014): Jurnal Teknologi Komputer dan Informatika
Publisher : Universitas Kristen Duta Wacana

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (3908.57 KB) | DOI: 10.21460/inf.2014.101.322

Abstract

Building futsal court need some planning, especially about the materials needed to build a futsal court such as synthetic grass, roofs, walls, benches, and some other materials. Each of materials have many choices usually based on quality and price. Thus it needs a system to assist on calculating the optimum materials combination based on a specified budget. Minimax Route method are used with dynamic programming techniques to maximize the quality of materials while minimizing the price of materials chosen. Based on system testing conducted to futsal court owners in Yogyakarta, the implementation are helpful and have many useful information for someone who want to build futsal court.
Implementasi Sistem Crowdsourced Labelling Berbasis Web dengan Metode Weighted Majority Voting Rachmat, Antonius; Lukito, Yuan
ULTIMA InfoSys Vol 6 No 2 (2015): UltimaInfoSys :Jurnal Ilmu Sistem Informasi
Publisher : Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (949.711 KB) | DOI: 10.31937/si.v6i2.223

Abstract

Crowdsourced Labelling is a large scale data labelling process, solicits a large group of people to label the data, usually via Internet.  This paper discusses about design and implementation of Web-based Crowdsourced Labelling.  Supervised learning classification methods need labelled training data for its training phase.  Unfortunately, in many cases, there aren’t any already available labelled training data.  Large scale data labelling is a tedious and time consuming work.  This research develops a web-based crowdsourced labelling which able to solicit a large group of people as data labeler to speed up the data labelling process.  This system also allows multiple labeler for every data.  The final label is calculated using Weighted Majority Voting method.  We grabbed and used Facebook comments from the two candidates’ Facebook Page of 2014 Indonesian Presidential Election as testing data.  Based on the testing conducted we can conclude that this system is able to handle all the labelling steps well and able to handle collision occurred when multiple labeler labelling a same data in the same time. The system successfully produces final label in CSV format, which can be processed further with many sentiment analysis tools or machine learning tools. Index Terms - Crowdsources labeling, web-based system, supervised learning, weighted majority voting.