Claim Missing Document
Check
Articles

Found 33 Documents
Search

IMPLEMENTASI PENGEMBANGAN METODE DIFFERENTIAL EVOLUTIONUNTUK CLUSTERING PIXEL Saikhu, Ahmad; Fahmi, Hisyam
JUTI: Jurnal Ilmiah Teknologi Informasi Vol 9, No 2, Juli 2011
Publisher : Department of Informatics, Institut Teknologi Sepuluh Nopember

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (277.541 KB) | DOI: 10.12962/j24068535.v9i2.a32

Abstract

Perkembangan metode komputasi telah mengalami percepatan yang luar biasa. Berbagai teknik komputasi untuk mendapatkan solusi dengan kinerja optimal terus berkembang. Sejumlah algoritma termasuk dalam rumpun Evolutionary Computation, diantaranya adalah Differential Evolution (DE) yang berhasil menyelesaikan masalah optimasi dalam berbagai bidang diantaranya masalah clustering. Keunggulan DE adalah karena implementasinya yangmudah dan kecepatan konvergensinya. Dalam clustering, DE menghadapi kendala penentuan jumlah cluster. Pada penelitian ini diimplementasikan sebuah algoritma Evolutionary Clustering (EC) yang merupakan pengembangan dari DE. EC diterapkan untuk melakukan pengelompokan pixel-pixel dari citra gray-scale atas beberapa area homogen yang berbeda satu dengan lainnya. EC tidak membutuhkan informasi awal tentang jumlah cluster yang akan terbentuk. EC menjadi salah satu solusi untuk menentukan jumlah cluster optimal dengan nilai validitas yang lebih baik. Kinerja dari EC akan dibandingkan dengan algoritma Fuzzy C-Means (FCM). Hasil dari EC dibanding FCM relatif sama dari segi nilai cluster validity index namun EC membutuhkan waktu relatif lebih singkat.
RANCANG BANGUN OPTIMASI KEBUTUHAN BAHAN BAKU MENGGUNAKAN ALGORITMA WAGNER-WHITIN Saikhu, Ahmad; Sarwosri, Sarwosri; Laila, Nur
JUTI: Jurnal Ilmiah Teknologi Informasi Vol 7, No 4, Juli 2009
Publisher : Department of Informatics, Institut Teknologi Sepuluh Nopember

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (239.01 KB) | DOI: 10.12962/j24068535.v7i4.a87

Abstract

Lotting or purchasing raw materials is one step in Material Requirement Planning. Lotting technique that already known is the Wagner-Within algorithm. This algorithm is widely used because it provides optimal solutions for problem sizedeterministic dynamic reservation at a particular time period in which the needs of the entire period must be completed. It takes a application system of optimization planning raw material requirements using the Wagner-Whitin algorithm. The development of this process begins with building a power module of demand data using Arima method (1,1,1), then followed by forecasting modules of consumer demand for end product by using the multiplicative decomposition forecasting methods, and ends with the development of Materials Requirement Planning module (MRP I) using the Wagner-Whitin algorithm. The results of the test system with test data is the generation of data will form the same pattern that is likely up from week to week. Forecasting results have high accuracy registration of 99.48%, 99.64% and 99.68%. Wagner-Whitin algorithm always produces the combination of weeks. Result of the combination in the first week will produces the minimum cost for the entire week of production.
Multivariate Time Series Forecasting Using Recurrent Neural Networks for Meteorological Data Hariadi, Victor; Saikhu, Ahmad; Zakiya, Nurotuz; Wijaya, Arya Yudhi; Baskoro, Fajar
SENATIK STT Adisutjipto Vol 5 (2019): Peran Teknologi untuk Revitalisasi Bandara dan Transportasi Udara [ISBN XXX-XXX-XXXXX-
Publisher : Sekolah Tinggi Teknologi Adisutjipto

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.28989/senatik.v5i0.365

Abstract

Rainfall is related to a number of factors that are interdependent and influenced by dynamic global time, region and climate factors. Determination of relevant predictors is important for the efficiency of the rainfall estimator model. Although some climate modeling studies in one region/country have high accuracy, this model is not necessarily suitable for other regions. Determination of predictor variables by considering spatio-temporal factors and local / global features results in a very large number of inputs. Feature selection produces minimal input so that it gets relevant predictor variables and minimizes variable redundancy. Recurrent Neural Networks is one of the artificial neural networks that can be used to predict time series data. This study aims to predict rainfall by combining the SVM classification method and the RNN method. Tests on the Perak 1 daily and monthly weather data (WMO ID: 96933) and Perak 2 Station daily and monthly data(WMO ID: 96937), showed high accuracy results with an R2 are 92.1%; 94.1%; 90.9% and 89.6%.
SEGMENTASI CITRA MENGGUNAKAN CLUSTERING DENGAN PENDETEKSI SADDLE POINT Saikhu, Ahmad; Soelaiman, Rully; Hambali, Imam
Proceedings of KNASTIK 2009
Publisher : Duta Wacana Christian University

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Segmentasi citra adalah proses membagi citra ke dalam region-region yang terpisah, di mana setiap region adalahhomogen dan mengacu pada sebuah kriteria keseragaman yang jelas. Segmentasi yang dilakukan terhadap citra harus tepatagar informasi yang terkandung di dalamnya dapat diterjemahkan dengan baik.Penelitian ini menggunakan algoritma meanshift untuk segmentasi. Mean shift merupakan prosedur nonparametric sederhana untuk mengestimasi kerapatan gradient.Metode ini mempunyai parameter untuk mengontrol resolusi hasil segmentasi. Untuk mendeteksi boundary klasterdigunakan algoritma saddle point.Berdasarkan hasil uji coba menunjukkan waktu proses dan jumlah klaster dalamsegmentasi tergantung pada nilai parameter bandwidth domain spasial hs, bandwidth domain range hr, dan klaster terkecil Myang dimasukkan. Semakin besar nilai hs maka waktu segmentasi semakin lama. Semakin besar nilai hs,hr, dan M makajumlah klaster semakin sedikit. Boundary yang dihasilkan saddle point menjadi lebih halus.
DISCRIMINANT ANALYSIS IMPLEMENTATION BASED ON VARIABLE PREDICTIVE MODELS FOR SIMILARITY PATTERN CLASSIFICATION Saikhu, Ahmad; Eka Putra, Deneng
Proceedings of KNASTIK 2012
Publisher : Duta Wacana Christian University

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

At present, there are many pattern classification methods that can be used such asLDA, kNN, Bayesian networks, CART, ANN and SVM. However, many classificationmethods mentioned above causes some issues. The problem are the large computationalcost, and weakness of methods mentioned above to classify the class because it is basedsolely on inter-class boundary (decision boundaries).As an alternative method otherthan the methods have already been exist, the relationship between features (interrelation)in a class can be used to classify a sample of a particular class. Based onthese ideas variable predictive model method based class discrimination (VPMCD) isproposed by (Raghuraj &Lakshminarayanan, 2008) as a new classification approach tothe problem of large data and overlapping which cannot be easily solved by the otherclassification methods.The testings are done using six well studied data sets (Diabetic,Hear, Iris, Wine, Digit, Letter). The results are equations wihich have capability toclasify new sample.
Performance Study Of Uncertainty Based Feature Selection Method On Detection Of Chronic Kidney Disease With SVM Classification Qolby, Lailly Syifa'ul; Buliali, Joko Lianto; Saikhu, Ahmad
IPTEK The Journal for Technology and Science Vol 32, No 2 (2021)
Publisher : IPTEK, LPPM, Institut Teknologi Sepuluh Nopember

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12962/j20882033.v32i2.10483

Abstract

Chronic Kidney Disease (CKD) is a disorder that impairs kidney function. Early signs of CKD patients are very difficult until they lose 25% of their kidney function. Therefore, early detection and effective treatment are needed to reduce the mortality rate of CKD sufferers. In this study, the authors diagnose the CKD dataset using the Support Vector Machine (SVM) classification method to obtain accurate diagnostic results. The authors propose a comparison of the result on applying the feature selec- tion method to get the best feature candidates in improving the classification result. The testing process compares the Symmetrical Uncertainty (SU) and Multivariate Symmetrical Uncertainty (MSU) feature selection method and the SVM method as a classification method. Several experimental scenarios were carried out using the SU and MSU feature selection methods using the CKD dataset. From the results of the tests carried out, it shows that using the MSU feature selection method with 80%: 20% data split produces nine important features with an accuracy value of 0.9, sensi- tivity 0.84, specification 1.0, and when viewed on the ROC graph, the MSU method graph shows the true positive value is higher than the false positive value. So the classification using the MSU feature selection method is better than the SU feature selection method by 90% accuracy
Implementasi Algoritma Rijndael dengan Menggunakan Kunci Enkripsi yang Berukuran Melebihi 256 bit Gracius Cagar Gunawan; Ahmad Saikhu; Rully Soelaiman
Jurnal Teknik ITS Vol 2, No 2 (2013)
Publisher : Direktorat Riset dan Pengabdian Masyarakat (DRPM), ITS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12962/j23373539.v2i2.3867

Abstract

Dalam dunia nyata, terdapat permasalahan keamanan informasi yang dapat diselesaikan dengan menggunakan enkripsi. Seiring dengan pertumbuhan teknologi, enkripsi dengan ukuran cipher key yang kecil semakin mudah dibongkar. Oleh karena itu, ukuran cipher key perlu ditingkatkan. Sampai saat ini, Advanced Encryption Standard yang dibentuk berdasarkan Algoritma Rijndael yang dapat menggunakan cipher key berukuran 256 bit masih dipakai. Dalam artikel ini, dilakukan studi dan implementasi algoritma enkripsi yang mampu menerima cipher key berukuran lebih dari 256 bit dengan bahasa C. Hasil uji coba menunjukkan program menghasilkan keluaran yang benar dan memiliki pertumbuhan waktu eksekusi secara linear, yaitu Q(Nb*Nk) dengan Nb adalah ukuran data masukan dan Nk adalah ukuran cipher key
Penjadualan Petugas Medis pada Kondisi Darurat dengan Menggunakan Binary Integer Programming Berbasis Web Bryan Alfadhori; Ahmad Saikhu; Victor Hariadi
Jurnal Teknik ITS Vol 5, No 2 (2016)
Publisher : Direktorat Riset dan Pengabdian Masyarakat (DRPM), ITS

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (464.318 KB) | DOI: 10.12962/j23373539.v5i2.17996

Abstract

Bencana alam maupun bencana yang disebabkan kelalaian manusia sering kali menimbulkan kondisi darurat. Penugasan petugas medis pada kondisi darurat merupakan hal yang sangat penting. Terbatasnya petugas dengan keahlian yang dibutuhkan yaitu kombinasi penugasan petugas medis yang tidak tepat dapat membuat penjadualan yang tidak optimal. Dalam penentuan petugas medis yang memenuhi kondisi, digunakan representasi graf bipartite dan algoritma Ford Fulkerson dalam proses untuk pemilihan petugas medis yang memenuhi kondisi tersebut. Binary integer programming digunakan untuk menentukan kombinasi penugasan yang optimal. Berdasarkan hasil uji coba dapat disimpulkan bahwa kedua proses yang diimplementasikan dapat membantu dalam pengambilan keputusan penugasan. Representasi dari graf bipartite terbukti dapat memberikan hasil yang akurat berupa petugas medis yang memenuhi kondisi. Binary Integer Programming juga memberikan hasil yang optimal berupa petugas medis yang ditugaskan dan total jarak yang paling minimal.
Implementasi Metode Integer Programming Untuk Penjadualan Tenaga Medis Pada Situasi Darurat Berbasis Aplikasi Mobile Ahmad Saikhu; Laili Rochmah
IC-ITECHS Vol 1 (2014): Prosiding IC-ITECHS 2014
Publisher : IC-ITECHS

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1926.19 KB)

Abstract

Unexpected events, including natural disasters and disasters caused by human negligence often results in the occurrence of an emergency condition. For a disaster event, disaster mitigation is not only necessary, but an emergency response plan for handling the health / safety of life of the victims, are very important. Proper medical treatment by a professional will save many lives. Such event can occur at any location and any time, and in the same time can occur a few emergency events that require some physician team with qualified doctors / particular specialization for handling a number of victims. On the other hand, physicians have a high mobility in performing his professional duties. Due to the limited number of physicians with certain qualifications as required in an emergency and demands should arrive quickly, we need a reliable system in order to coordinate the assignment of an emergency event can be handled properly. Coordination and scheduling can be done by those who have an authority. Scheduling optimization model with special characteristics, among others the existence of a number of emergency events at several locations at the same time, the limited number of physician, needs special skills for handling emergencies, and the presence of medical personnel who are mobile can be done by developing one assignment optimization model, namely integer linear programming (ILP). As for fast communication for coordination between the management authority with a number of doctors, a mobile application can be applied. In this research, Bipartite Graph (BG) model, Ford-Fulkerson Algorithm (FFA) and IPL model as the main part of the decision making structure are implemented into a program. The first model (BG) present concordance between physicians with their expertises. The second model (FFA) helps the authority to determine physician who meet the condition. The third model (ILP) helps the authority to determine the best combination of assignment physician. Input, process, output, and feedback from the implementation is represent decision making system for the assignment physician in to the emergency events. So that the physicians can quickly find out the assignment, the schedule can be accessed through mobile applications.
PERBAIKAN SISTEM STANFORD RTE PADA KALIMAT MENGANDUNG EKSPRESI ARITMATIKA Rakhmat Arianto; Daniel Oranova Siahaan; Ahmad Saikhu
Lontar Komputer : Jurnal Ilmiah Teknologi Informasi Vol. 4, No. 2 Agustus 2013
Publisher : Institute for Research and Community Services, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (112.959 KB)

Abstract

Sistem Stanford Recognizing Textual Entailment adalah sistem yang dapat mendeteksi entailment maupun kontradiksi yang terkandung dalam pasangan kalimat Text dan Hypothesis.Pada tahun 2009, dilakukan penelitian pengembangan sistem Stanford RTE dengan menggabungkan sistem Stanford RTE pada tahun 2006 dengan sistem Stanford RTE pada tahun 2008.Salah satu kelemahan pada sistem Stanford RTE tahun 2009 adalah kesalahan deteksi kontradiksi pada pasangan kalimat Text dan Hypothesis mengandung ekspresi aritmatika.Untuk menutupi kelemahan tersebut, ditambahkan fitur aritmatika dalam sistem Stanford RTE yang khusus memproses pasangan kalimat Text dan Hypothesis mengandung ekspresi aritmatika.Fitur aritmatika dibangun dengan empat tahap utama, yaitu tahap analisa linguistik, tahap pencarian nilai kemiripan kata, tahap penentuan operator aritmatika, dan tahap penyimpulan entailment, kontradiksi, atau tidak diketahui.Fitur aritmatika telah dilakukan pengujian terhadap 30 pasang kalimat Text dan Hypothesis mengandung ekspresi aritmatika yang diambil dari halaman website berita dengan hasil tingkat keberhasilan mencapai 80%.