AIUB Journal of Science and Engineering (AJSE)

A Comparison of Customer Churn Vector Embedding Models with Deep Learning

2024-04-25T21:56:01+06:00

In the telecommunication industry, Deep learning has been utilized for churn prediction. Some companies have used sophisticated deep learning techniques to predict churn, which yielded good results. However, future studies are still required to evaluate several deep learning mechanisms as only SoftMax Loss has been used so far. By comparing customer churn vector embedding models with several methods, including SoftMax Loss, Large Margin Cosine Loss, Semi-Supervised Learning, and a combination of Large Margin Cosine Loss and Semi-Supervised Learning, we continue our previous research to apply deep learning in predicting customer churn in the telecommunications industry in this paper. The use of Large Margin Cosine Loss has been proven in face recognition which can increase the discrimination between vectors embedding in different classes. Understanding how mixing unlabeled and labeled input might alter developing algorithms and learning behavior that benefit from this combination are the goals of semi-supervised learning. This approach successfully encouraged feature discrimination in customer behavior as well as improved the overall accuracy of the model. Large Margin Cosine Loss in this study achieved 83.74% of the F1 Score compared to other methods. It was further demonstrated that the produced vectors for churn prediction are discriminative by examining the cluster's similarity and the t-SNE plot.

GCN-Net: 3D Point Cloud Classification & Localization Using Graph-CNN

2024-04-25T21:55:44+06:00

In this paper, we have demonstrated the application of a graph convolutional neural network for the purpose of object detection in a LiDAR point cloud. In order to encode the point cloud in the most time-effective manner, we make use of a near-neighbors graph with a defined radius. We create a graph convolutional neural network so that we can find out what kind of object and what class each vertex in a graph represents. We design a box merging and scoring operation to reliably combine detections from numerous vertices into a single score, and we offer an auto-registration strategy as a means of reducing the amount of translation errors that occur inside the system. According to the results of our tests using the KITTI benchmark, we are able to draw the conclusion that the method that was suggested achieves competitive accuracy with the point cloud, even beating fusion-based methods in some instances. According to the results of our research, the graph neural network has the potential to become an effective new tool for the detection of 3D objects.

Simulated Channel Length Variation Effects on Regulated Cascode Input Stage-Based Transimpedance Amplifier for Fiber Optics Applications

2024-04-25T21:55:29+06:00

A proposed transimpedance amplifier with channel length variation is simulated. The amplifier consists of a regulated cascode input stage followed by a common gate-common source configuration. A channel length series (45 nm, 90 nm and 130 nm) in CMOS technology was introduced within the proposed amplifier in order to achieve comparative performance analysis. There are two key findings from this study. On one hand, it was found that the trade off in gain versus bandwidth and input referred noise current still applies when channel length is moved upward from 45 nm up to 130 nm. A series of transimpedance amplifier gains (42.16 dBΩ, 44.34 dBΩ and 46.25 dBΩ) that correspond to (1.80 GHz, 1.33 GHz and 1.06 GHz) of bandwidths is reported corresponding to the above channel length series respectively with an input referred noise current spectral density series (16.35 pA/sqrt(Hz), 12.17 pA/sqrt(Hz) and 10.60 pA/sqrt(Hz)) of reduction. On the other hand, a reduction in power consumption occurred as channel length is moved upward for the same proposed topology. A total power consumption series (0.611 mW, 0.287 mW and 0.173 mW) was reported that corresponds to the above channel length series.

WVEHDD: Weighted Voting based Ensemble System for Heart Disease Detection

2024-04-25T21:55:11+06:00

Although several machine learning (ML) based algorithms are proposed by various researchers for Heart Disease detection (HDD), most of these works considered a very small experimental dataset to justify the efficiency of ML techniques in HDD. Moreover, despite of the low correlation of the features with the target, all the features were used for HDD. Considering the limitations of these existing systems, current study emphasizes on the designing of a Weighted Voting based Ensemble (WVE) Classifier for HDD from a sufficiently large dataset comprising of 1296 instances. Although there are 13 features, only 4 features are found to be statistically significant in HDD. For designing an efficient WVE classifier for HDD, the weighted votes of five efficient classifiers are combined to get the final decision. The experimental result shows that the proposed WVEHDD system outperforms the existing systems by providing the highest train accuracy of 96.15% and test accuracy of 95.64%

Performance Prediction of A Power Generation Gas Turbine Using An Optimized Artificial Neural Network Model

2024-04-25T21:54:55+06:00

This paper presents the application of an Artificial Neural Network (ANN) based model for performance ‎prediction of a power generation gas turbine. The suggested model was optimized to provide a large ‎database for comparison between different ANN topologies. Then, based on the optimization results, the ‎Multi-Layer Perceptron (MLP) of two layers was constructed and utilized for this study as the best-‎optimized topology. Training of this model was done using historical operational data of a Rolls Royce ‎‎(RB21-24G) gas turbine unit. The outcome results from this model used for performance prediction show ‎good accuracy for different ambient conditions and variable power ratings. Then, a degradation study was ‎also introduced comparing measurements of the same gas turbine utilizing one year later, on-site ‎operational data, with the predicted values generated by the ANN model. The result shows consistency ‎between the measured data and the model results.‎

Performance Analysis of Automatic Generation Control for a Multi-Area Interconnected System Using Genetic Algorithm and Particle Swarm Optimization Technique

2024-04-25T21:54:37+06:00

The primary focus of this paper is to assess an interconnected power system using different optimization techniques. The main purpose is to employ different optimization techniques, including genetic algorithms (GA) and particle swarm optimization (PSO), to systematically enhance the performance of a multi-area or two-area automatic generation control (AGC) system, aiming to optimize the three PID controllers gain values and improve system performance under diverse loading conditions. Two case studies are conducted exploring different loading conditions in the megawatt (MW) range, including increasing load demand and decreasing load demand. The analysis involves four scenarios, covering without any kind of controller, another with solely a proportional integral derivative (PID) controller, a PID controller enhanced through a genetic algorithm (GA), and lastly, a PID controller improved through particle swarm optimization (PSO). The optimization process utilizes the integral time absolute error (ITAE) as the objective function to evaluate the system's performance. The simulation outcomes for ITAE, settling time, overshoot, and undershoot for frequency deviation of area one, area two, and power deviation in the tie-line are compared with previous similar studies to assess the novelty of this work. The article highlights the importance of the multi-area AGC system and the significance of different optimization techniques in improving its performance.

Addressing Environmental Challenges in Technology Adoption within the Digital Supply Chain of Readymade Garments

2024-04-25T21:54:22+06:00

The study aims to investigate the challenges of technology adoption in the Digital Supply Chain (DSC) of the Readymade Garment (RMG) business, focusing on environmental barriers. The study utilizes the T-O-E framework to examine a sample of 380 participants, consisting of owners and top managers from RMG facilities. The data is analyzed using PLS-SEM Modeling with the aid of SmartPLS4 software. It examines the complex interrelationships between competitive pressure, consumer, external support, stakeholder networks, environmental concerns, and technology adoption within the digital supply chain of the RMG industry. The technology adoption is substantially influenced by environmental conditions, particularly regarding the alignment with customers. The analysis of mediation sheds light on the significance of the environment in both partial and complete mediation, as it exerts influence on competitive pressures, customer involvement, external support, and stakeholder networks. Comprehending this interrelationship is crucial for making well-informed business and policy formulation decisions. organizations must incorporate environmental factors into their strategic decision-making processes, ensuring sustainable technologies are adopted. Policymakers can employ these findings to implement environmentally sustainable policies, promoting innovation within the RMG sector. These measures guarantee the long-term viability of the industry and promote ecological accountability.

Advancing Fuzzy Logic: A Hierarchical Fuzzy System Approach

2024-04-25T21:54:06+06:00

Fuzzy logic systems (FLS) are widely used in various engineering, medical, and scientific applications for modelling complex and uncertain systems. However, traditional FLS has limitations in handling complex and hierarchical structures due to their lack of scalability and interpretability. This paper proposes an approach to hierarchical fuzzy systems (HFS) that enhances the traditional FLS by providing a hierarchical structure with multiple levels of fuzzy rules. The main contribution of this paper is the proposal of HFS, which improves interpretability, scalability, and accuracy compared to traditional FLS, particularly for real-world applications. However, the question arises, "How can the FLS be converted into the HFS?" In this paper, the approach to HFS architecture will comprise two levels of FLS, where the first level determines the overall behaviour of the system, and the second level refines the output by considering the local behaviour. The proposed approach has been validated through experimental results on a case studies, such as the Iris flower classification. The results demonstrate that HFS provides more efficient and reliable solutions and can be applied to various complex and hierarchical systems in different domains, such as manufacturing, robotics, and decision-making.

Predictions of Malaysia Age-Specific Fertility Rates using the Lee-Carter and the Functional Data Approaches

2024-04-25T21:53:49+06:00

Global fertility has been experiencing a significant decline, reaching towards the replacement ratio. This trend, coupled with increasing life expectancies, has led to the emergence of an ageing population. In this study, we delve into an analysis of fertility patterns among Malaysian women, considering both their childbearing age and ethnic groups. A comprehensive 63-year fertility dataset, from 1958 to 2020, were obtained from the Department of Statistics Malaysia. These data were fitted into the Lee-Carter model and its modified version, which is the functional data model. The models were evaluated using the out-sample forecast error measures. Results indicate that the third-order functional data model able to capture most of variation present in the actual data, consequently outperforming the Lee-Carter model in forecasting fertility rates among Chinese and Indian populations. The 20-year forecasts reveal a noteworthy shift in maternal ages of the highest births to older ages suggesting a trend towards delayed pregnancies among women. It is predicted that the Malay total fertility rates will likely fall to below the replacement level reaching 1.71 in 2040 whereas Chinese and Indian total fertility rates will substantially decrease to the lowest level in history below 1.0 which are 0.54 and 0.70 respectively. The evolution in Malaysian fertility rates is an alarming fact as, together with low mortality rates, it may impact the Malaysian population structure in future. Proactive policy measures are urgently needed to address these demographic shifts.

Performance Analysis of Non-Realistic Routing Protocol using Random Waypoint Model in MANET

2024-04-25T21:53:31+06:00

A mobile ad hoc network (MANET) is a collection of wireless nodes connected via wireless networks and has no set structure. MANETs feature a self-organized routing topology in which mobile nodes are free to move, making it difficult and crucial to construct a stable and reliable network. Thus, failure of the route is also regarded as a prime factor affecting the efficiency of any MANET routing protocol. The breaking of the connection between two routes or more nodes will cause the failure of the route specifically in the non-realistic routing protocol. In a network of mobile nodes, the link break is mainly based on the mobility of individual nodes. Therefore, the objective of this research is to investigate the performance of proactive DSDV and reactive AODV routing protocol using the Random Waypoint(RWP) mobility model in MANET. NS-2 network simulator is used to simulate the MANET environment and BonnMotion is to create a movement of mobile nodes that integrate with the routing protocol. The network performance metrics used are throughput, packet delivery ratio, and average end-to-end delay. In addition, three simulation scenarios have been conducted to compare AODV and DSDV routing protocols with varying numbers of nodes, a comparison of AODV and DSDV routing protocols with varying pause time, and a comparison of AODV and DSDV routing protocols with varying mobility speed. The result from the three scenarios analysed and concluded that the RWP mobility model with AODV gives a better performance of throughput with 869.69 kbps and Packet Delivery Ratio (PDR) with 83.00% meanwhile, RWP with DSDV is better for the average end-to-end delay(EED) with 212.970 bps.