Computer Engineering Commons^™

การทำนายแนวโน้มของตลาดหุ้นด้วยแบบจำลองการเรียนรู้เชิงลึกที่เพิ่มประสิทธิภาพร่วมกับข้อมูลเชิงตัวเลขและข้อมูลเชิงตัวอักษร, พิศุทธ อ่อนเจริญ

Chulalongkorn University Theses and Dissertations (Chula ETD)

การทำนายแนวโน้มของตลาดหุ้นเป็นสิ่งที่ยากเนื่องจากตลาดหุ้นมีความผันผวนสูงและได้รับอิทธิพลจากปัจจัยภายนอกอื่น ๆ ในปัจจุบันเริ่มมีการนำเทคนิคการเรียนรู้เชิงลึก (Deep Learning) มาใช้ในการทำนายแนวโน้มของตลาดหุ้น โดยที่ข้อมูลรับเข้าของแบบจำลองสามารถแบ่งออกได้เป็น 2 ประเภทคือ 1) ข้อมูลเชิงตัวเลข เช่น ราคาในอดีตและตัวชี้วัดทางเทคนิค และ 2) ข้อมูลเชิงตัวอักษร ซึ่งได้แก่หัวข้อข่าวและเนื้อข่าว เป็นต้น แต่อย่างไรก็ตามงานวิจัยส่วนใหญ่มักจะมุ่งเน้นไปที่สร้างแบบจำลองโดยใช้ข้อมูลประเภทใดประเภทหนึ่งเท่านั้น ในขณะที่นักลงทุนส่วนใหญ่ทำการวิเคราะห์พฤติกรรมของตลาดโดยพิจารณาจากข้อมูลหลากหลายประเภท งานวิจัยนี้ได้นำเสนอแบบจำลองการเรียนรู้เชิงลึกที่สามารถพิจารณาข้อมูลทั้งสองประเภทเพื่อทำนายแนวโน้มของตลาดหุ้น ซึ่งแบบจำลองนี้ประกอบไปด้วยนิวรอลเน็ตเวิร์กแบบคอนโวลูชัน (Convolutional Neural Network) และหน่วยความจำระยะสั้นแบบยาว (Long-Short Term Memory) โดยใช้ข้อมูลรับเข้าเป็นเหตุการณ์ฝังตัวซึ่งสกัดได้จากหัวข้อข่าว ราคาในอดีตและตัวชี้วัดทางเทคนิคซึ่งสร้างจากข้อมูลของราคาในอดีต รวมทั้งได้ทำการนำเสนอฟังก์ชันวัตถุประสงค์ชนิดใหม่ที่สามารถช่วยเพิ่มประสิทธิภาพในด้านผลตอบแทนเฉลี่ยต่อปีที่ได้จากการจำลองการซื้อขาย โดยการนำเอาค่าชาร์ปเรโชซึ่งเป็นตัวชี้วัดผลตอบแทนเมื่อเทียบกับความเสี่ยงมาใช้ร่วมกับค่าครอสเอนโทรปี

Integrated Reward Scheme And Surge Pricing In A Ride Sourcing Market, Hai Yang, Chaoyi Shao, Hai Wang, Jieping Ye

Research Collection School Of Computing and Information Systems

Surge pricing is commonly used in on-demand ride-sourcing platforms (e.g., Uber, Lyft and Didi) to dynamically balance demand and supply. However, since the price for ride service cannot be unlimited, there is usually a reasonable or legitimate range of prices in practice. Such a constrained surge pricing strategy fails to balance demand and supply in certain cases, e.g., even adopting the maximum allowed price cannot reduce the demand to an affordable level during peak hours. In addition, the practice of surge pricing is controversial and has stimulated long debate regarding its pros and cons. In this paper, to address the …

A Proposal For A High Availability Architecture For Voip Telephone Systems Based On Open Source Software, Alejandro Martin, Eric Gamess, Dedaniel Urribarri, Jesús Gómez

Research, Publications & Creative Work

The inherent needs of organizations to improve and amplify their technological platform entail large expenses with the goal to enhance their performance. Hence, they have to contemplate mechanisms of optimization and the improvement of their operational infrastructure. In this direction arises the need to guarantee the correct operation and non-degradation of the services provided by the platform during the periods with a significant load of work. This type of scenario is perfectly applicable to the field of VoIP technologies, where users generate elevated loads of work on critical points of the infrastructure, during the process of interaction with their peers. …

Efficient Information Dissemination In Vehicular Networks With Privacy Protection, Xiaolu Cheng

Theses and Dissertations

Vehicular ad hoc network (VANET) is a key component of intelligent transportation System (ITS). In VANETs, vehicles and roadside units exchange information for the purpose of navigation, safe driving, entertainment and so on. The high mobility of vehicles makes efficient and private communications in VANETs a big challenge.

Improving the performance of information dissemination while protecting data privacy is studied in this research. Meet-Table based information dissemination method is first proposed, so as to improve the information dissemination, and to efficiently distribute information via utilizing roadside units, Cloud Computing, and Fog Computing. A clustering algorithm is proposed as well, to …

Improving The Performance And Energy Efficiency Of Emerging Memory Systems, Yuhua Guo

Theses and Dissertations

Modern main memory is primarily built using dynamic random access memory (DRAM) chips. As DRAM chip scales to higher density, there are mainly three problems that impede DRAM scalability and performance improvement. First, DRAM refresh overhead grows from negligible to severe, which limits DRAM scalability and causes performance degradation. Second, although memory capacity has increased dramatically in past decade, memory bandwidth has not kept pace with CPU performance scaling, which has led to the memory wall problem. Third, DRAM dissipates considerable power and has been reported to account for as much as 40% of the total system energy and this …

Consuming Digital Debris In The Plasticene, Stephen R. Parks

Theses and Dissertations

Claims of customization and control by socio-technical industries are altering the role of consumer and producer. These narratives are often misleading attempts to engage consumers with new forms of technology. By addressing capitalist intent, material, and the reproduction limits of 3-D printed objects’, I observe the aspirational promise of becoming a producer of my own belongings through new networks of production. I am interested in gaining a better understanding of the data consumed that perpetuates hyper-consumptive tendencies for new technological apparatuses. My role as a designer focuses on the resolution of not only the surface of the object through 3-D …

A Distributed Intrusion Detection Model Via Nondestructive Partitioning And Balanced Allocation For Big Data, Xiaonian Wu, Chuyun Zhang, Runlian Zhang, Yujue Wang, Jinhua Cui

Research Collection School Of Computing and Information Systems

There are two key issues in distributed intrusion detection system, that is, maintaining load balance of system and protecting data integrity. To address these issues, this paper proposes a new distributed intrusion detection model for big data based on nondestructive partitioning and balanced allocation. A data allocation strategy based on capacity and workload is introduced to achieve local load balance, and a dynamic load adjustment strategy is adopted to maintain global load balance of cluster. Moreover, data integrity is protected by using session reassemble and session partitioning. The simulation results show that the new model enjoys favorable advantages such as …

Real-Time Inbound Marketing: A Use Case For Digital Banking, Alan Megargel, Venky Shankararaman, Srinivas K. Reddy

Research Collection School Of Computing and Information Systems

Over the years banks have been strategically using digital technologies to help transform various aspects of their business. In recent times, this strategy has evolved into one of digital augmentation of the bank’s processes, products and channels. This allows for reaching out to customers and partners through digital platforms, for example; the addition of mobile apps for customers to access and perform service transactions. Marketing continues to play a major role in supporting the expansion of business and increasing revenue for the bank. Marketing has evolved from mass direct targeting to more personalised, face-to-face and real-time targeting. In the digital …

The Measurement Of Shielding Effectiveness For Small-In-Size Ferrite-Based Flat Materials, İsa Araz

Turkish Journal of Electrical Engineering and Computer Sciences

This work presents the results of our studies on the electromagnetic interference shielding effectiveness of small-in-size samples of ferrite composition microwave absorber prepared by using the ceramic method. The shielding effectiveness measurement was performed using the complex intrinsic parameters of the material. The coaxial holder method with the well-known test technique based on transmission/reflection measurements was used in the experimental setup. Complex intrinsic parameters were measured in the frequency range of 2-18 GHz by using the Nicholson - Ross - Weir (NRW) method with an automated test setup. First, the shielding effectiveness was measured by conventional test method and then …

Emi Filter Design Based On The Separated Electromagnetic Interference In Switched Mode Power Supplies, Samet Yalçin, Şükrü Özen, Selçuk Helhel

Turkish Journal of Electrical Engineering and Computer Sciences

Usage of switch mode power supplies (SMPS) has been increased tremendously in recent years due to their advantages compared to conventional ones. In spite of their advantages, SMPSs cause conducted and radiated emissions due to the fact that they switch on and off at specific frequencies. Electromagnetic interference (EMI) filters are mostly preferred equipment for the reduction of conducted emission for coupled circuits. Improved EMI filter in the literature to suppress both common and differential mode noises sourced by ATX (a sample of SMPS) power supply has been proposed. A proposed filter was designed by use of AWR Microwave Office …

Classification And Regression Analysis Using Support Vector Machine For Classifying And Locating Faults In A Distribution System, Sophi Shilpa Gururajapathy, Hazlie Mokhlis, Hazlee Azil Bin Illias

Turkish Journal of Electrical Engineering and Computer Sciences

Various fault location methods have been developed in the past to identify the faulty phase, fault type, faulty section, and distance. However, this identification is commonly conducted in a separate manner. An effective fault location should be able to identify all of these at the same time. Therefore, in this work, a method using a support vector machine (SVM) to identify the fault type, faulty section, and distance considering the faulty phase is proposed. The proposed method uses voltage sag magnitude of the distribution system as the main feature for the SVM to identify faults. The fault type is classified …

The Impact Of Transmission Power Levels Set Size On Lifetime Of Wireless Sensor Networks In Smart Grids, Hüseyi̇n Uğur Yildiz

Turkish Journal of Electrical Engineering and Computer Sciences

Wireless sensor networks (WSNs) have been confirmed as one of the most promising technologies for many smart grid (SG) applications due to their low complexity and inexpensive costs. A typical WSN is formed with numerous battery limited sensor nodes mounted on critical components of a SG system for monitoring applications. Acquired monitoring data by sensor nodes are conveyed to the base station generally by using multihop communication techniques. WSN-based SG applications encounter severe propagation losses due to extreme channel conditions of the SG environment. In order to reduce possible packet errors caused by channel variations, transmission power control approaches can …

Management Of A Hybrid Renewable Power Plant Supplying An Isolated Rural Load Within A Changing Environment, Hichem Azli, Mohamed Mostefai

Turkish Journal of Electrical Engineering and Computer Sciences

Often, authors deal with the sizing issue of hybrid power plants on horizons of several years. Proposed approaches are certainly essential to optimize energy costs. However, these solutions cannot remain optimal throughout the life cycle of the plant due to the inevitable evolution of the number of households, their consumption profiles, and possible degradation of a part of the plant. In this paper, an efficient management strategy of a sized hybrid renewable system is developed. It is based on the scheduling of the different resources of the plant. The main aim is to minimize the generated energy cost while ensuring …

Improved Design Of Axial Flux Permanent Magnet Generator For Small-Scale Wind Turbine, Mojtaba Eldoromi, Sajjad Tohidi, Mohammad Reza Feyzi, Naghi Rostami, Reza Emadifar

Turkish Journal of Electrical Engineering and Computer Sciences

Recently, much attention has been given to axial flux permanent magnet generators due to some advantages such as high efficiency, high power density, and higher torque compared to radial flux generators. In addition, increased number of poles makes them appropriate for low-speed applications such as wind power generators. In this study, an axial flux permanent magnet synchronous generator with unique features such as high power density is designed for small-scale wind turbines. The structure of generator includes a rotor and a stator. The generator is designed and then analyzed by Flux 11.2 software. The analysis includes the effects of air …

Average Modeling And Evaluation Of 18-Pulse Autotransformer Rectifier Unit Without Interphase Transformers, Shahbaz Khan, Xiaobin Zhang, Husan Ali, Haider Zaman, Muhammad Saad, Bakht Muhammad Khan

Turkish Journal of Electrical Engineering and Computer Sciences

This paper presents an improved average model and evaluation of an 18-pulse autotransformer rectifier unit (ATRU) in differential delta configuration. Average models remove the switching behavior of diode rectifiers and high bandwidth transients, which not only facilitates simulation of power systems by reducing computational cost but also enables impedance-based stability analysis for large complex power systems. To experimentally validate the proposed average model, a 2-kW, 18-pulse ATRU has been tested and the results of the derived model are compared with those of the switching model and experimental prototype. Computed transfer function of load impedance from the proposed average model closely …

Application Of Acf-Wavelet Feature Extraction For Classification Of Some Artificial Pd Models Of Power Transformer, Vahid Parvin Darabad

Turkish Journal of Electrical Engineering and Computer Sciences

In this paper, 7 different artificial partial discharge (PD) models of power transformer defect types are built in a high voltage laboratory, and then PD signals are recorded in the whole power frequency cycle to develop a preliminary PD knowledge base. A specific technique is used to extract single PD events from the recorded knowledge base. By application of wavelet transform, different decomposition levels of extracted PD signals are investigated. Feature extraction is performed by application of statistic moments on the autocorrelation function of PD signal decompositions. "Best" features are selected based on the F-test. To evaluate the performance of …

Humanoid Robots Supporting Children’S Learning In An Early Childhood Setting, Helen Crompton, Kristen Gregory, Diane Burke

Teaching & Learning Faculty Publications

This qualitative study explored the affordances provided by the integration of the NAO humanoid robot in three preschool classrooms. Using the Head Start Early Learning Outcomes Framework as a lens, the researchers qualitatively analyzed data from focus groups, observations, field notes and student artifacts, using grounded coding to uncover language and communication, physical, cognitive and social–emotional learning experiences for children. The researchers also examined interactions between the robot, children and teachers to identify successes and challenges experienced during the integration. Findings indicate the robot provided opportunities for student development in all learning domains. Students were intellectually curious about the robot; …

Collaborative Fall Detection Using Smartphone And Kinect, Xue Li, Lanshun Nie, Hanchuan Xu, Xianzhi Wang

Research Collection School Of Computing and Information Systems

Humanfall detection has attracted broad attentions as sensors and mobile devices are increasingly adopted in real-life scenarios such as smart homes. The complexity of activities in home environments pose severe challenges to the fall detection research with respect to the detection accuracy. We propose a collaborative detection platform that combines two subsystems: a threshold-based fall detection subsystem using mobile phones and a support vector machine (SVM)-based fall detection subsystem using Kinects. Both subsystems have their respective confidence models and the platform detects falls by fusing the data of both subsystems using two methods: the logical rules-based and D-S evidence fusion …

Generating Adversarial Attacks For Sparse Neural Networks, Jack H. Chen, Walt Woods

Maseeh Summer Undergraduate Research Experience

Neural networks provide state-of-the-art accuracy for image classification tasks. However traditional networks are highly susceptible to imperceivable perturbations to their inputs known as adversarial attacks that drastically change the resulting output. The magnitude of these perturbations can be measured as Mean Squared Error (MSE). We use genetic algorithms to produce black-box adversarial attacks and examine MSE on state-of-the-art networks. This method generates an attack that converts 90% confidence on a correct class to 50% confidence of a targeted, incorrect class after 2000 epochs. We will generate and examine attacks and their MSE against several sparse neural networks. We theorize that …

User-Friendly Deniable Storage For Mobile Devices, Bing Chang, Yao Cheng, Bo Chen, Fengwei Zhang, Wen-Tao Zhu, Yanju Liu, Zhan Wang

Research Collection School Of Computing and Information Systems

Mobile devices are prevalently used to process sensitive data, but traditional encryption may not work when an adversary is able to coerce the device owners to disclose the encryption keys. Plausibly Deniable Encryption (PDE) is thus designed to protect sensitive data against this powerful adversary. In this paper, we present MobiPluto, a user-friendly PDE scheme for denying the existence of sensitive data stored on mobile devices. A salient difference between MobiPluto and the existing PDE systems is that any block-based file systems can be deployed on top of it. To further improve usability and deniability of MobiPluto, we introduce a …

Improvement Of Heart Attack Prediction By The Feature Selection Methods, Hi̇dayet Takci

Turkish Journal of Electrical Engineering and Computer Sciences

Prediction of a heart attack is very important since it is one of the leading causes of sudden death, especially in low-income countries. Although cardiologists use traditional clinical methods such as electrocardiography and blood tests for heart attack prediction, computer aided diagnosis systems that use machine learning methods are also in use for this task. In this study, we used machine learning and feature selection algorithms together. Our aim is to determine the best machine learning method and the best feature selection algorithm to predict heart attacks. For this purpose, many machine learning methods with optimum parameters and several feature …

Estimating Left Ventricular Volume With Roi-Based Convolutional Neural Network, Feng Zhu

Turkish Journal of Electrical Engineering and Computer Sciences

The volume of the human left ventricular (LV) chamber is an important indicator for diagnosing heart disease. Although LV volume can be measured manually with cardiac magnetic resonance imaging (MRI), the process is difficult and time-consuming for experienced cardiologists. This paper presents an end-to-end segmentation-free method that estimates LV volume from MRI images directly. The method initially uses Fourier transform and a regression filter to calculate the region of interest that contains the LV chambers. Then convolutional neural networks are trained to estimate the end-diastolic volume (EDV) and end-systolic volume (ESV). The resulting models accurately estimate the EDV and ESV …

Automatic Detection Of The Respiratory Cycle From Recorded, Single-Channel Sounds From Lungs, Seli̇m Aras, Mehmet Öztürk, Ali̇ Gangal

Turkish Journal of Electrical Engineering and Computer Sciences

Listening to the sounds made by the lungs is a long-standing method that is still used to diagnose lung diseases. Many studies have been conducted on the automatic recognition of recorded sounds from the lungs. However, in these studies, respiratory cycles, i.e. inhalations followed by exhalations, were either monitored manually or by using multichannel signals of the sounds made by the lungs. In order to recognize sounds made by the lungs automatically, one must first use these sounds to determine the respiratory cycles. Our previous study was the first study presented in the literature in which the boundaries of respiratory …

A Novel Multistage System For The Detection And Removal Of Pectoral Muscles Inmammograms, İdi̇l Işikli Esener, Semi̇h Ergi̇n, Tolga Yüksel

Turkish Journal of Electrical Engineering and Computer Sciences

In this paper, a novel multistage scheme for pectoral muscle removal from mammography images is proposed, and the performance of this system is verified using the publicly available Mammographic Image Analysis Society digital mammogram database. This database is composed of mediolateral oblique mammography images including three different tissue types (fatty, fatty-glandular, and dense-glandular) with three health status types (normal, benign cancer, and malignant cancer). In the implementation of the proposed system, a mammography image is first preprocessed by performing noise reduction background removal followed by artifact suppression processes. Then a presegmentation procedure is applied using region growing and line fitting …

Multimodal Medical Imaging Repository And Information System In Epilepsysurgery, Gülsüm Akdeni̇z, İbrahi̇m Atli, Fati̇h Vehbi̇ Çelebi̇

Turkish Journal of Electrical Engineering and Computer Sciences

Surgery is an option for patients with drug-resistant epilepsy, but it requires a comprehensive assessment. Electroencephalography (EEG), magnetic resonance imaging (MRI), and functional MRI (fMRI) are used to localize the epileptogenic zone, which directly affects the surgery outcome. Accessing EEG, MRI, and fMRI results and patient information simultaneously using traditional methods might result in misinformation and increase the workload of clinicians. In this study, we developed a modern web-based repository system for the preoperative evaluation of epilepsy disorder, including multimodal medical images and patient information. Our dedicated system is enriched with clinical metadata that are not currently available and is …

Gnss Augmentation Through Ku-Band Communication Satellites With Rtkrealization, Şenol Gülgönül

Turkish Journal of Electrical Engineering and Computer Sciences

In this study, a method to broadcast real-time kinematics (RTK) corrections over a communication satellite is proposed. RTK corrections generated by the reference station are uplinked to the communication satellite within a multiplexed DVB-S carrier. Corrections are received by a satellite dish and the DVB-S satellite receiver. Better than 18 cm of accuracy is achieved for rover distances of up to 500 km in Turkey. The proposed method can be used as an alternative satellite-based augmentation system for Turkey and other countries.

Ultra Low-Power Dc Voltage Limiter For Rfid Application In 0.18-Μm Cmostechnology, Rezvan Dastanian, Sajad Nejadhasan

Turkish Journal of Electrical Engineering and Computer Sciences

In this paper, a low-power DC voltage limiter is designed for radio frequency identification tags. In this design, the low-power bandgap reference (BGR) circuit, which is insensitive to temperature, and also the comparator are used for the voltage limiter circuit. To make the I-V curve approach the ideal one, four inverter stages are employed to the output of the comparator. The structure of the BGR and comparator use only MOSFET transistors that work in subthreshold region when the limiter is inactive. These two blocks have power consumption of 220 pW and 1300 pW, respectively. The limited output voltage is 1.5 …

A Strategy For Improving Netclust Server Placement For Multicloud Environments, Yuting Zhao, Jinhong Liu, Qing Fang, Lifang Xu, Changlai Du, Yuan Xiaoqun

Turkish Journal of Electrical Engineering and Computer Sciences

In this paper, we propose a fast server placement algorithm to improve the NetClust framework and make it more efficient and exible. To this end, we introduce hierarchical clustering technologies to the NetClust framework and propose a exible server placement algorithm, which integrates the agglomerative and divisive clustering technologies to reduce the time complexity and avoid the performance uctuation affected by the initial node selection. The experiment results show that our server placement algorithm may reduce the time complexity of server selection of NetClust significantly and improve the exibility and applicability of the NetClust.

Performance Analysis Of Hamming Code For Wsn-Based Smart Grid Applications, Meli̇ke Yi̇ği̇t, Vehbi̇ Çağri Güngör, Pinar Bölük

Turkish Journal of Electrical Engineering and Computer Sciences

Many methods have been employed to detect, compare, and correct errors to increase communication reliability and efficiency in wireless sensor networks (WSNs). However, to the best of our knowledge, no existing study has compared the performance of error control codes by using different modulation techniques in a smart grid communication environment when multichannel scheduling is used. This paper presents a detailed performance evaluation and makes a comparison of different modulation techniques, such as frequency shift keying (FSK), differential phase shift keying (DPSK), binary phase shift keying (BPSK), and offset quadrature phase-shift keying (OQPSK), using Hamming codes in a 500-kV line-of-sight …

Energy-Efficient And Reliable Data Collection In Wireless Sensor Networks, Deepa Puneeth, Nishanth Joshi, Pradeep Kumar Atrey, Muralidhar Kulkarni

Turkish Journal of Electrical Engineering and Computer Sciences

Ensuring energy efficiency, data reliability, and security is important in wireless sensor networks (WSNs). A combination of variants from the cryptographic secret sharing technique and the disjoint multipath routing scheme is an effective strategy to address these requirements. Although Shamir's secret sharing (SSS) provides the desired reliability and information-theoretic security, it is not energy efficient. Alternatively, Shamir's ramp secret sharing (SRSS) provides energy efficiency and data reliability, but is only computationally secure. We argue that both these approaches may suffer from a compromised node (CN) attack when a minimum number of nodes is compromised. Hence, we propose a new scheme …