Claim Missing Document
Check
Articles

Implementasi Metode Learning Vector Quantization Untuk Klasifikasi Penyakit Demam Desiani, Nurhidayati; Muflikhah, Lailil; Dewi, Candra
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 3 No 2 (2019)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (422.175 KB)

Abstract

Fever is an early symptom of various diseases that have been experienced by almost everyone. Some of the diseases include typhoid fever, malarial fever and dengue fever. These three diseases have similar early symptoms. Similar symptoms of each disease often cause difficulty in obtaining anamnese (temporary diagnosis) so that patients get the initial handling is less precise and further worsen the condition of the patient. To overcome this required a system that can facilitate in identifying the disease based on the symptoms felt by the patient. In this study using Learning Vector Quantization method which is a method of classification. The system works with the training and testing phases that will result in classes of typhoid fever classes, malarial fever and dengue fever. The parameters used are 15 parameters of symptoms of febrile illness. The best average accuracy result is 100% using comparison of test data and training data of 10:90, learning rate 0,1, learning rate reduction constant 0,1, minimum learning rate 10-5, and maximum number of iteration 10.
Klasifikasi Kualitas Susu Sapi Menggunakan Metode Support Vector Machine (SVM) Sari, Puspita; Muflikhah, Lailil; Wihandika, Randy Cahya
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 3 (2018)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1057.665 KB)

Abstract

Cow milk has a lot of animal protein and have benefit for children and whoever in process for grow up. Cow milk contains good essential amino acids. Malang Animal Health Laboratory as the unit executor in east java Animal Husbandry Department do a test in kesmavet for efforts to secure milk as a farm product with appropriate testing in suitable with the Indonesian National Standard (SNI). The classification of cow milk quality is still using organoleptic (smell, taste, color) that are linguistic, so that variable and parameter are uncertain and become themain obstacle of expert in determining good milk quality. To resolve this issue, this can be done with schizophrenia classification using support vector machine (SVM) algorithm, which SVM performace is more suitable than other classification methods. In this study there are 269 data that is divided into two data that is data training and data testing with three classification result, that is low, medium, and hight. The result in this paper get the best acuracy based ratio data 50%:50%, with Kernel RBF and λ (lambda) = 0,001, C (complexity) = 0,01, γ (gamma) =0,00001, maximum iteration = 30 and σ kernel RBF= 2. The average result of accuracy using SVM method in cow milk quality classification was 92.82% and highest accuracy was 94.02%.
Pengenalan Entitas Bernama untuk Identifikasi Transaksi Akuntansi Menggunakan Hidden Markov Model Rizqiyah, Rika Raudhotul; Muflikhah, Lailil; Fauzi, Mochammad Ali
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 7 (2018)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (906.78 KB)

Abstract

Accounting is a task which has an important role in supporting economic continuity, due to the recording of any business process that occurred was done in accounting. However, the recording of financial transactions in accounting for identification into journal is still done manually, so that required classification and extraction of information contained in the accounting transaction text to make it easier. Named Entity Recognition (NER) is the first step needed to perform information extraction. To solve this problem, named entity recognition done for identification of accounting transaction. In this research used method of Hidden Markov Model (HMM), because HMM can resolve labeling task and and known robustly in performing named entity recognition. The main process in this named entity recognition is divided into modeling process using Hidden Markov Model and decoding process using Viterbi Algorithm. In this research will be recognize 12 entities namely DATE, TITLE, PER, TRANS, EXP_MON, TYP_COMP, FIRST_ORG, SECOND_ORG, EXP_DATE, NO_DATE, MONTH and YEAR. Overall entity recognition with addition Laplace Smoothing and Regular Expression techniques produce a value of average precision, recall and f-measure consecutive 81.75%, 87.88%, and 82.39%.
Pengembangan Metode Klasifikasi Berdasarkan K-Means Dan LVQ Eka Ratnawati, Dian; ., Marji; Muflikhah, Lailil
Jurnal Teknologi Informasi dan Ilmu Komputer Vol 1, No 1 (2014)
Publisher : Fakultas Ilmu Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (646.659 KB)

Abstract

AbstrakPada penelitian ini dikembangkan metode klasifikasi berdasarkan pengelompokan K-Means dan LVQ. Metode-metode klasifikasi yang telah ada jika ada data dengan frekuensi kecil cenderung tidak digunakan dalam pengujian kelas, padahal dimungkinkan data tersebut sangat bermanfaat. Langkah untuk melakukan pengelompokan adalah: melakukan pengelompokkan dengan K-Means. Pengelompokan terus dilakukan sampai mencapai threshold (batasan tertentu). Jika threshold sudah dicapai dan pada satu cluster masih terdapat kelas yang berbeda maka dilakukan pembelajaran dengan menggunakan LVQ. Akurasi gabungan K-Means dan LVQ lebih baik daripada dengan K-Means murni. Untuk  akurasi rata-rata tertinggi K-Means dan LVQ didapatkan 92%, sedang untuk K-Means murni 82%.Kata kunci: klasifikasi, pengelompokan, K-Means, LVQAbstractThis research will develop methods of classification based on K-Means clustering. Grouping method used is a combination of K-Means and LVQ. Classification methods that have been there if there is a small frequency data tend to be used in the test class, but it is possible they are very useful. Steps to perform grouping is doing the K-Means clustering. Grouping is continues until it reaches the threshold. If the threshold has been reached and there are cluster of different classes then performed using LVQ learning. Accuracy combined K-Means and LVQ is better than with pure K-Means. For the highest average accuracy of K-Means and LVQ gained 92%, while for the K-Means only 82%.Keywords: classification, grouping, K-Means, LVQ
Prediksi Nilai Mata Kuliah Mahasiswa Menggunakan Algoritma K-Apriori Muflikhah, Lailil; Yunita, W. Lisa; Furqon, M. Tanzil
SISFO Vol 6 No 2 (2017)
Publisher : Department of Information Systems, Institut Teknologi Sepuluh Nopember

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

Abstract

Tujuan utama dari prediksi nilai mata kuliah adalah membantu mahasiswa mengambil mata kuliah pilihan secara tepat. Kebanyakan mahasiswa mengambil mata kuliah didasarkan pada jumlah mahasiswa mengambil matakuliah. Sekumpulan transkrip mahasiswa dapat dianalisis pola keterkaitan (association rule) antar nilai matakuliah. K-Apriori merupakan metode data mining untuk mencari pola keterkaitan nilai mata kuliah sehingga dapat digunakan memprediksi nilai mata kuliah lain. Tahapan utama metode ini meliputi mengelompokkan data menggunakan metode K-Means dan menemukan pola nilai mata kuliah menggunakan Apriori. Namun terdapat kekosongan nilai karena seluruh mata kuliah yang ditawarkan tidak diambil setiap mahasiswa. Oleh karenanya, dilakukan preprocessing data menggunakan Wiener Transformation sebelum dicari polanya.Pengujian didasarkan tingkat kemampuan akademik mahsiswa dengan minimum support dan confidence sebesar 10% dan lift ratio >1. Hasilnya, rule yang dibangkitkan dari IPK di bawah dan di atas rata-rata memiliki tingkat kesalahan sebesar 8.75% dan 8.5%. Sedangkan jika rule dibangkitkan dari IPK rata-rata memiliki kesalahan sebesar 11%.
Optimasi Variasi Menu Makanan Sesuai Gizi Pada Anak Panti Asuhan Dengan Improved Particle Swarm Optimization Caturrahmanto, Aldino; Muflikhah, Lailil; Cholissodin, Imam
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 3 No 9 (2019)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (991.576 KB)

Abstract

Nutrition is a compound which important to human growth and their health. Nutrition consumption is really important that they can affect health if poorly managed. Unfortunately, there are still places which have poorly managed nutrition planning, which one of them is Orphanage. Improved Particle Swarm Optimization is an Evolutionary Algorithm which is based on nature folks of birds flying in group searching for new food points. This Algorithm have great ability to search local optimum solution and also global optimum solution. In this case, a list of foods menu will be represented by a Particles which consisting indexes of number represent menu. The result of our testing found that swarm size of 70 and combination of 2,0 for C1 and C2 is the best parameter for this problem. Although giving great solution based on Nutrition, this algorithm still offer total price above our limit value.
Implementasi Metode Fuzzy Subtractive Clustering Untuk Pengelompokan Data Potensi Kebakaran Hutan/Lahan Kusuma, Vianti Mala Anggraeni; Muflikhah, Lailil; Furqon, Muhammad Tanzil
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 1 No 9 (2017)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1258.645 KB)

Abstract

Abstract Forest is the habitat for all kinds of animals and plants, forests have a very big function to maintain the balance of nature, as the supplier of the oxygen requirement for living on earth, and the natural resources that provide a variety of materials for human needs. But at this moment the existence of forest diminishing due to illegal logging by humans or by forest fires are becoming more frequent. Forest fires this gives very bad impact, extinction of some species of plants and animals, the smoke is detrimental to health even low and so forth. So to be able to help deal with the issue made a system that can manage data hotspots (hotspots) with Fuzzy subtractive clustering. Parameter data used in the development of the system: brightness temperature and FRP (Fire Radiative Power). The result of clustering which illustrates the potential of forest fires, which are grouped in the high potential and low potential. The test results showed the best coefficient silhouette value of 0.45 and the results of the cluster is formed by two clusters using radius values ​​0.2, accept ratio 0.5, reject ratio 0.15. The results of the analysis in the determination of the potential for forest fires result is a high potential with an average brightness value of 335.727⁰K, FRP 57.248 and average confidence 83.47%. While medium potential with an average brightness value of 318.934⁰K, FRP 23.330 and average confidence 58.08%. Keywords: Clustering, Hotspot, Fuzzy subtractive clustering, Silhouette Coefficient
Klasifikasi Berita Online dengan menggunakan Pembobotan TF-IDF dan Cosine Similarity Herwijayanti, Bening; Ratnawati, Dian Eka; Muflikhah, Lailil
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 1 (2018)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1034.548 KB)

Abstract

In discussing the online news by using the weighting of tf-idf and cosine of this similarity the previous research reference on online news information using single pass clustering algorithm, where the data to be used comes from the online news website that is kompas.com. Because of the many news that is on the website, so sometimes the news is posted not in accordance with the category. Human error will be the problem of wrong news posting. In addition to posting errors online news groupings are also important for the convenience of users to search for news according to their category. Implementing online news stories using tf-idf and cosine similarities, preprocessing processes ie tokenizing, stopword and stemming can reduce the term process of speeding the weighting of terms using tf-idf and accelerating the cosine process of similarity. The goal is to facilitate human error as well as reduce caution categorization. The value is able to classify news with accreditation rate of 91.25%.
Implementasi Metode Simple Additive Weighting (SAW) Untuk Penentuan Penerima Zakat Prayogo, Hanggar Wahyu Agi; Muflikhah, Lailil; Wijoyo, Satrio Hadi
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 11 (2018)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (466.292 KB)

Abstract

Zakat can be one way done in effort to increase social welfare and was one of the important role in economic community empowerment. Other benefits of zakat is to help reduce poverty and also can reduce the social inequality in society. But the problems encountered by these institutions in distributing zakat is the inaccuracy in choosing a recipient of zakat. The inaccuracy in choosing a zakat recipient occurs because the zakat management agency still provides subjective assessments. It is causing loss to the people who are more eligible to receive the zakat. The SAW (Simple Additive Weighting) is one of the methods in decision support system that can be used to create a system that can determine the recipient of zakat. There are 4 criteria used in determining zakat recipient in this research, there are family status, family income, total dependents and value report card. Based on the accuracy testing that has been done by using 60 test data, obtained the best accuracy of 90%. The SAW method can be applied properly in determining the recipient of zakat.
Klasifikasi Risiko Hipertensi Menggunakan Metode Neighbor Weighted K-Nearest Neighbor (NWKNN) Yudha, Bayu Laksana; Muflikhah, Lailil; Wihandika, Randy Cahya
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 2 (2018)
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (850.713 KB)

Abstract

Hypertension or called high blood pressure caused high risk of death in Indonesia. This could trigger sustainable effect to other diseases such as heart attack and kidney failure. According to WHO, as many as 30% of Indonesians are sufferers hypertension, Indonesian Hypertension Doctor Association also said that 76% of cases of hypertension can not diagnosed earlier and therefore hypertension is called a silent killer. The way to handling hypertension earlier is by early detection of hypertension in form of Early Alertness System (SKD). In this research will classified risk of hypertension based on medical record using Neighbor Weigted K-Nearest Neighbor (NWKNN) classification method. This method is the development of the KNN method. In NWKNN there is a weighting process in each class of hypertension risk. In this study, the classification of hypertension into 4 risks that is Normal, Pre Hypertension, Stage 1 and Stage 2. The results of this research shows that the NWKNN method is able to classify the hypertension risk well when tested on 100 training data, 25 testing data, K score=10, and E score=4 with accuracy result that reached 88%.
Co-Authors Agung Setiyoaji, Agung Agus Ardiansyah, Agus Andrianto, Rheza Raditya Ardhani, Luthfi Afrizal Attabi', Ahmad Wildan Bachtiar, Fitra Abdurrachman Bayu Laksana Yudha, Bayu Laksana Bayu Rahayudi Bening Herwijayanti, Bening Bhaskara, Kukuh Bhuana, Ksatria Bonita, Olivia Candra Dewi Caturrahmanto, Aldino Desiani, Nurhidayati Dian Eka Ratnawati Dian Eka Ratnawati Edy Santoso Eva Agustina Ompusunggu, Eva Agustina Faris Dinar Wahyu Gunawan, Faris Dinar Wahyu Hamim Fajar, Muh Harahap, Eni Hartika Haryanto, Dimas Joko Imam Cholissodin Indriati Indriati Lestari, Puji Indah M. Tanzil Furqon, M. Tanzil Marji . Marji Marji Mochammad Ali Fauzi, Mochammad Ali Muhammad Abduh Muhammad Fajri Muhammad Tanzil Furqon, Muhammad Tanzil Munawarah, Robbiyatul Nugraha, Randi Pratama Nurul Dyah Mentari, Nurul Dyah Nurul Hidayat Pamungkas, Vidya Capristyan Prayogo, Hanggar Wahyu Agi Pristiyanti, Ria Ine Puspita Sari Putri, Gessia Faradiksi Rachmad Indrianto, Rachmad Randy Cahya Wihandika Ratih Kartika Dewi Rekyan Regarsari Mardhi Putri, Rekyan Regarsari Mardhi Rekyan Regasari Mardi Putri, Rekyan Regasari Mardi Rendi Cahya Wihandika, Rendi Cahya Rika Raudhotul Rizqiyah, Rika Raudhotul Rizal Setya Perdana Rizby, Laily Putri Royyan, Ahmad Nur Safitri, Annisaa Amalia Sasmita, Aldi Bagus Satrio Hadi Wijoyo Septian, Ardiza Dwi Shidiq, Brillian Ghulam Ash Sigit Adinugroho Suprapto Suprapto Surya Dermawan, Surya Suwandy, Yogi Syafruddin Agustian Putra, Syafruddin Agustian Syahputra, Dhimas Wida Tampubolon, Yobel Leonardo Tibyani Tibyani, Tibyani Utami, Tahtri Nadia Vianti Mala Anggraeni Kusuma, Vianti Mala Anggraeni Wahyu Rizki Ferdiansyah, Wahyu Rizki Yulian Ekananta, Yulian Yuliana, Marine Putri Dewi Yunita, W. Lisa