Hasil Pencarian  ::  Simpan CSV :: Kembali

"ABSTRAKPerforma baik dari Self-Organizing Map (SOM) telah terbukti dalammengklasifikasikan citra wajah yang berada dalam kondisi pencahayaan yangbaik. Namun saat objek wajah mengalami pencahayaan yang berubah-ubah dandiambil dari berbagai sudut pandang berbeda, maka tingkat nilai rekognisi citrawajah dengan menggunakan metode SOM umumnya akan menurun.Dalam penelitian ini menggunakan metode Fuzzy Self-Organizing Map(FSOM) sebagai sistem pengenal wajah pada citra untuk meningkatkan nilairekognisi citra wajah yang mengalami pencahayaan yang berubah-ubah dariberbagai sudut.Hasil penelitian menunjukkan bahwa performa FSOM untuk mengenali wajahpada data berdasarkan sudut pandang dengan tingkat rekognisi tertinggididapatkan pada set data ke-10 saat sudut 100 pada saat wajah frontal yaitusebesar 87%, pada data berdasarkan perubahan cahaya dengan tingkat rekognisitertinggi pada set data ke-1 sebesar 66.88%, dan pada data berdasarkan objekwajah dengan tingkat rekognisi tertinggi pada set data ke-4 sebesar 88.33%.Berdasarkan Hasil penelitian didapatkan bahwa tingkat rekognisi rata-rata FSOM30% lebih tinggi dari SOM pada setiap pengelompokkan data dan juga didapatkanbahwa dengan metode FSOM mampu mengenali citra dengan baik yangmengalami pencahayaan yang berubah-ubah dari sudut yang berbeda-beda.

---

ABSTRACTPerform of Self-Organizing Map (SOM) has been proven to classify the face

images in good illumination conditions. But when this technique is applied to

various viewpoints of images in unstable illumination conditions, the accuracy

of face recognition will decrease.

In this research, Fuzzy Self-Organizing Map (FSOM) is introduced as a new

technique to increase the accuracy when the images are taken from various

viewpoints in the change illumination conditions.

In this results from the research show that perform of FSOM to face

recognition from based on the viewpoints have the highest recognition rate in the

tenth data set when the viewpoints is set to be 10 degree where the images had

been taken is achieved at 87%. The result has also shown that based on the

illumination conditions, the highest recognition rate is achieved at 66.88% in the

first data set. Based on the face objects, the most accurate recognition is achieved

at 88.33% in the fourth data set. These results show that FSOM can give 30%

better performance than SOM to perform face images classification in the changes

illumination conditions and various viewpoints."

"ABSTRAK"Metode pengendalian suatu proses dengan menggunakan pengendali fuzzy selama ini telah berkembany dengan pesat Perkembangan ini disebabkan oleh kemampuan pengendali fuzzy mengatasi ketidaklinieran suatu sistem Pengendali fuzzy yang dapat mengimplernentasikan pola pikir dari seorang operator dapat menghasilkan pengendaii yang ""cerdas"". Hal mii tidak dapat dihasilkan oleh pengendali yang biasanya didesain secara matematis dan mengahaikan beberapa sifat sistem yang tidak linier agar didapatkan pengendali yang sederhana.NaMun, dalam mendesain suatu pengendali fuzzy pun terdapat beberapa masalah seperti tidak mudahnya membentuk suatu fungsi keanggotaan input dan output pengendali, sulitnya menentukan aturan yang dipergunakan sebagal rule base. Kesulitan menentukan aturan ini karena pada dasarnya tiap ahli (expert) akan mengutarakan aturan-aturan yang berbeda.Untuk itulah diperlukan tuning aturan atau faktor skala agar diperoleh unjuk kerja pengendali yang baik. Ketika pengendali didesain, pengendali di-tune sesuai dengan keadaan simulasi sistem. Bahkan jika pengendali telah di-tune dengan balk, faktor skala atau aturan tersebut perlu di-tune setelah instalasi pengendali.Steam generator pada pembangkit listrik tenaga nuklir adalah sistem yang penting karena faktor keamanan pembangkit listrik dan is merupakan sistem yang menghasilkan uap bagi turbin yang akan menggerakan generator dan kemudian akan menghasilkan listrik. Karena hal-hal tersebut seseorang tidak dapat melakukan tuning secara terus menerus untuk menghindari kecelakaan atau efek lain yang tidak diinginkan karena proses tuning yang tidak tepat."

---

"

"The theme of this book is to examine the feasibility of creating such robots within the limitations of current mechanical engineering. The topics comprise the following aspects of such a pursuit : the philosophy of design of self-organizing mechanical systems, self-organization in biological systems, the history of self-organizing mechanical systems, a case study of a self-assembling/self-repairing system as an autonomous distributed system, a self-organizing robot that can create its own shape and robotic motion, implementation and instrumentation of self-organizing robots, and the future of self-organizing robots. All topics are illustrated with many up-to-date examples."

"The book provides answers by developing a general principle, homeokinesis, the dynamical symbiosis between brain, body, and environment. That is shown to drive robots to self- determined, individual development in a playful and obviously embodiment- related way, a dog-like robot starts playing with a barrier, eventually jumping or climbing over it, a snakebot develops coiling and jumping modes, humanoids develop climbing behaviors when fallen into a pit, or engage in wrestling-like scenarios when encountering an opponent. The book also develops guided self-organization, a new method that helps to make the playful machines fit for fulfilling tasks in the real world.The book provides two levels of presentation. Students and scientific researchers interested in the field of robotics, self-organization and dynamical systems theory may be satisfied by the in-depth mathematical analysis of the principle, the bootstrapping scenarios, and the emerging behaviors. But the book additionally comes with a robotics simulator inviting also the non- scientific reader to simply enjoy the fabulous world of playful machines by performing the numerous experiments."

"How do minds work? In Exploring Robotic Minds: Actions, Symbols, and Consciousness as Self-Organizing Dynamic Phenomena, Professor Jun Tani reviews key experiments within his own pioneering neurorobotics research project aimed at answering this fundamental and fascinating question. The book shows how symbols and concepts representing the world can emerge via deep learning within robots, by using specially designed neural network architectures by which, given iterative interactions between top-down proactive subjective and intentional processes for plotting action, and bottom-up updates of the perceptual reality after action, the robot is able to learn to isolate, to identify, and even to infer salient features of the operational environment, modifying its behavior based on anticipations of both objective and social cues. Through permutations of this experimental model, the book then argues that longstanding questions about the nature of consciousness and freewill can be addressed through an understanding of the dynamic structures within which, in the course of normal operations and in a changing operational environment, necessary top-down/bottom-up interactions arise. Written in clear and accessible language, this book opens a privileged window for a broad audience onto the science of artificial intelligence and the potential for artificial consciousness, threading cognitive neuroscience, dynamic systems theory, robotics, and phenomenology through an elegant series of deceptively simple experiments that build upon one another and ultimately outline the fundamental form of the working mind."

"Bank yang menyelenggarakan program keuangan inklusif cenderung tidak melakukan diferensiasi pelayanan bagi para pelanggannya yang berasal dari populasi masyarakat unbanked. Banyak organisasi mengganggap segmen keuangan inklusif sebagai ladang yang kurang populer dalam mendapatkan keuntungan, karena kecilnya pemasukan dan besarnya biaya operasional yang dibutuhkan. Namun, beberapa studi sebelumnya tentang keuangan inklusif dan segmentasi pelanggan telah menentang gagasan ini dan menyatakan bahwa klasifikasi lebih lanjut terhadap kategori pelanggan khusus ini dapat membawa keuntungan bagi pihak bank. Segmentasi pelanggan sering dilakukan menggunakan model Recency, Frequency, dan Monetary (RFM) untuk mendapatkan nilai pelanggan bagi perusahaan.Penelitian ini bertujuan untuk membuat model prediksi segmen pelanggan dalam konteks keuangan inklusif, dengan menerapkan penggunaan dua teknik data mining Self-Organizing Map (SOM) dan C5.0 decision tree (DT) secara hybrid. Analisa sosioekonomi, regional, dan pengeluaran digunakan untuk menilai pelanggan, alih-alih menggunakan RFM. Penelitian ini juga mempelajari pengaruh SOM terhadap kinerja klasifikasi keseluruhan, yang dievaluasi menggunakan confusion matrix. Dataset yang digunakan memiliki struktur generik sehingga model ini diharapkan dapat membantu pengembangan program keuangan inklusif pada institusi keuangan penyelenggara keuangan inklusif lainnya.

---

Banks adopting financial inclusion program often exclude differentiation in their services towards the target customers. Many organizations consider financial inclusion inflicts huge operational costs hence it is deemed infamous for profit gain. Previous studies in financial inclusion and customer segmentation have challenged this notion, concluding that further classification of this particular customer class could indeed bring profit for the bank and such that maintaining existing profiting customers induce less cost than the effort of acquiring new customers. Customer segmentation is often done using the Recency, Frequency and Monetary (RFM) model to assess a customer's value for the company.

This study aims to model customer segment predictions in the context of financial inclusion, using socioeconomic, regional, and expenditure analyses to assess customer values. Two data mining techniques Self-Organizing Map (SOM) and C5.0 decision tree (DT) are used in a hybrid setting. This study also observes the effect of SOM on overall classification performance, which is evaluated using confusion matrix. Due to the generality of the input dataset, the prediction model is expected to be usable, with minimal adjustments, by other financial inclusion institutions in need of customer segmentation."

"Penelitian ini terdiri dari dua tahap. Growing self-organizing map (GSOM) algorithm dan hybrid bee colony optimization (BCO) dan self-organizing map (SOM) untuk mengimprove SOM performance. Pada tahap pertama GSOM digunakan untuk menentukan SOM topology dan pada tahap kedua, hybrid BCOSOM digunakan untuk mengadjust SOM weights. Metode BCOSOM akan dibandingkan dengan metode PSO, BCO, SOM, PSOSOM, SOM+PSO, dan SOM+BCO dengan menggunakan 4 benchmark data sets (Iriss, Glass, Wine, dan Vowel). Dari hasil komputasi menunjukkan bahwa metode BCOSOM dapat mencari solusi yang lebih baik dari algoritma lainnya. Dari hasil tersebut, BCOSOM digunakan pada Group Technology untuk menentukan part families pada komponen plat disebuah perusahaan medical furniture di Yogyakarta.

---

This research proposes a two stage method growing self organizing map GSOM algorithm and bee colony optimization BCO based self organizing map BSOSOM to improve SOM performance. In the first stage GSOM is used to determine the SOM topology and then followed by BCOSOM to fine tune the SOM weights. The proposed BCOSOM algorithm is compared with other algorithms PSO BCO SOM PSOSOM SOM PSO and SOM BCO using four benchmark data sets Iris Glass Wine and Vowel. The computational result indicates that BCOSOM algorithm is able to find a better solution than other algorithms. Furthermore, the proposed algorithm has been also employed to Group Technology to cluster components into part families for a medical manufacture in Indonesia."

"Penggunaan gas bumi yang merupakan non-renewable energy, perlu lebih diefisienkan. Hal tersebut menyebabkan segmentasi pelanggan gas di sektor industri penting untuk dilakukan, agar dapat dibentuk strategi pemasaran atau penetapan tarif yang tepat. Penelitian ini dilakukan di PT Perusahaan Gas Negara (PGN) dengan menggunakan salah satu metode data mining, yaitu Self-Organizing Map (SOM), untuk dihasilkan klasterisasi pelanggan berdasarkan karakteristik penggunaannya, sebagai acuan dalam pembentukan segmentasi pelanggan gas bumi. Variabel area, jenis pelanggan, sektor industri, rata-rata penggunaan, standar deviasi penggunaan, dan total penyimpangan menjadi variabel yang digunakan dalam penelitian ini. Hasilnya adalah terdapat 37 klaster dan 9 segmen yang terbentuk, dari 838 data pelanggan yang digunakan dalam penelitian ini. Kesembilan segmen ini menggambarkan secara umum karakteristik pelanggan gas bumi di PT PGN.

---

The usage of the natural gas which is non-renewable energy, needs to be more efficient. It also results the customer segmentation is the necessary thing to do, in order to set up a marketing strategy or a determination of the appropriate tariff. This research was conducted at PT PGN using one of the methods of data mining, i.e. Self-Organizing Map (SOM), that resulted the clustering of customer based on the characteristic of its user, as a reference to create the customer segmentation of natural gas user. Variable of area, type of customer, the industrial sector, the average usage, standard deviation of the usage, and the total deviation become the variable which are used in this research. It results 37 cluster and 9 segment, from 838 customer data which are used in this research. These 9 segments illustrate the general characteristic of the natural gas customer of PT PGN."