Welcome to CTU Journal of Innovation and Sustainable Development

EMD combined with ensemble of machine learning predictors for foreign exchange rate forecasting

Duong Tuan Anh*, Tran Van Xuan

Abstract |

PDF

Forecasting foreign exchange rates is a critical financial challenge. In this paper, we build on recent trends and address the limitations of prior research by proposing a novel approach. Our method combines empirical mode decomposition (EMD) with ensemble of machine learning predictors in foreign exchange rate forecasting. To demonstrate that our proposed method (called EMD-ML) is effective, we used the new approach to forecast six foreign exchange rate time series at a specific time. The first experiment was implemented to compare the proposed forecasting model EMD–LSTM, which combines empirical mode decomposition (EMD) with ensemble of Long Short-Term Memory (LSTM) models, and the single LSTM model. The results indicate that the proposed EMD–LSTM model is more effective than the single LSTM. Besides, to aim at comparing deep-learning models against shallow machine learning models in combination with the EMD decomposition, the second experiment compared EMD-LSTM with the approach which combines EMD with an ensemble of k-nearest neighbors’ predictors (called EMD-KNN method) and the results of the second experiment show that EMD-LSTM cannot outperform EMD-KNN in foreign exchange rates forecasting.

Demonstration of Grover’s algorithm for retrieving secret keys in a basic SPN block cipher

Vu Minh Thang, Nguyen Van Nghi*, Le Quoc Dat, Do Quang Trung

Abstract |

PDF

In this study, we present in detail the application of Grover's quantum algorithm to the searching problem of the secret key of a simple SPN (Substitution–permutation network) block cipher called Yo-yo. The main goal of the paper is to clarify the construction of the quantum circuit and the operation phases of Grover's algorithm to find the secret key with the condition of knowing at least 1 pair of plaintext-ciphertext. To achieve this goal, we consider 2 cases: the case where there is a unique key that satisfies and the case where there are 2 keys that satisfy at the same time. As a result, our implementation technique, implemented in the Qiskit programming language, requires only 17 qubits to find the key of the Yo-yo block cipher correctly. This technique can be effectively applied on IBM quantum computers for large-scale SPN block ciphers, such as AES and GOST R.34.10.2015, which are widely used today.

Using U-Net models in deep learning for brain tumor detection from MRI scans

Nguyen Minh Khiem*, Tran Phuoc Huy, Phan Tan Tai

Abstract |

PDF

Tumor diseases in the nervous system are both dangerous and complex. Magnetic Resonance Imaging (MRI) is crucial for detecting brain disease; however, identifying the presence of tumors from these is time-consuming and requires a professional doctor. Utilizing deep learning for tumor detection in MRI images can reduce waiting times and enhance detection accuracy. We propose a method employing two U-Net models: ResNeXt-50 and EfficientNet architectures, integrated with a Feature Pyramid Network (FPN) for segmenting brain tumor. The models were trained on the BraTS 2021 dataset, consisting of 3,929 MRI scan images with 3,929 corresponding masks, divided into training, testing, and evaluation sets in a 70:15:15 ratio. The results indicate that the hybrid model, which combines EfficientNet and FPN, delivers superior performance, with an average Intersection over Union (IoU) accuracy of 0.90 on the test set compared to 0.50 for ResNeXt-50, and Dice accuracy of 0.92 compared to 0.66 for ResNeXt-50. Furthermore, we developed a web application that implements the EfficientNet with FPN model, facilitating convenient tumor detection from uploaded MRI images for doctors.

Assessing wind energy exploitation potential in several regions of Viet Nam using Kernel density estimation model

Tin Trung Chau, Tuan Ngoc Nguyen, Ton Duc Do*

Abstract |

PDF

This article analyzes and assesses the potential for wind energy exploitation in six regions of Vietnam. The wind speed data are used to construct wind speed probability distributions (WSPDs) based on kernel density estimation (KDE). The KDE distribution, with six bandwidth selection methods, is implemented to generate probability density functions (PDFs) for each region's data to describe wind speed characteristics. The statistical tests Cramér-Von Mises (CvM), Anderson-Darling (A-D), and Kolmogorov-Smirnov (K-S) are applied to evaluate the PDFs' goodness-of-fit performance. The analysis results present the KDE distribution using the least-squares cross-validation (LSCV), and the Scott bandwidth selection method has outstanding fitting performance. Based on these PDF distributions, the wind turbine (WT) power curve is used to estimate and predict the amount of electricity that can be produced. This study also proposes a reliable method for wind power output planning based on wind speed that can be universally applied.

Students' performance prediction employing Decision Tree

Abtahi Ahmed, Farzana Akter Nipa, Wasi Uddin Bhuyian, Khaled Md Mushfique, Kamrul Islam Shahin, Huu-Hoa Nguyen*, Dewan Md. Farid

Abstract |

PDF

An optimized educational community is a must in this modern era. The intersection of educational activities and the transformative potentials of Educational Data Mining (EDM) should be traversed, highlighting the reasoning behind the importance of EDM. Prior prediction of how a student stands academically, can facilitate them towards a much safer approach with their life decisions. This study uses the vast power and analytical domain of EDM, combining it with machine learning models, upholding an accurate prediction of students' academic performance. The study consists of a dataset containing academic, demographic and social data of undergraduate students. The paper aims to analyze comprehensively the features that act behind academic performance. Lastly, it compares the impact of non-academic data separately on a student's performance and with academic data as well. Traditional machine learning algorithms perform quite well in general, with SVM giving a best accuracy of around 95% with academic data, while training and testing the model without academic data still gives a good performance of 93%. The hierarchical tree from Decision Tree visualizes the key features, which include past results, family members' qualification levels and their jobs, hobbies of the student, commute time, and more.

An automated data collection process for constructing graph data relying on LLMs

Ho Ngoc Ton, Nguyen Hoang Son, Nguyen Ngoc Minh Chau, Pham-Nguyen Cuong*

Abstract |

PDF

This paper introduces a process that is designed to harvest data automatically from a variety of online sources. The core of this process lies in its data-handling techniques, which include drawing, cleaning, deduplicating, extracting, and categorizing of raw data to convert unstructured data into a structured format represented and imported in a graph database. The data extraction step utilizes Large Language Model (LLMs) for Named Entity Recognition (NER). A case study on deploying course data collection illustrates the enhancements brought about by this automation, showcasing improvements in the accuracy, completeness, and timeliness of updates in the course data. An evaluation carried out on the extraction and matching methods shows that the F1-score and precision rates are high. Overall, this study contributes to advancement of the field by providing a methodology for automating the collection and processing of online data sources, significantly improving the quality of data collection from online sources.

Application of deep learning for rice leaf disease detection in the Mekong Delta

Ngo Duc Luu*, Le Thi Thuy Diem, Ha Thi Phuong Anh

Abstract |

PDF

The Mekong River Delta, the largest rice-producing region in Vietnam with an annual output of over 25 million tons, plays a vital role in ensuring food security both within the country and globally. In recent years, it has undergone significant transformation in rice cultivation, which aims to support farmers here to plant rice more effectively. However, severe weather conditions and soil degradation have negatively impacted rice growth. Additionally, rice is highly susceptible to various diseases that must be identified and prevented promptly. As a result, leveraging technology such as AI and deep learning to diagnose rice diseases based on leaf symptoms is essential. This paper utilizes an image dataset of three common rice leaf diseases—leaf smut, brown spot, and bacterial leaf blight—and applies deep learning networks (MobileNet and ResNet) to evaluate and select the best model. A diagnostic program is then developed to detect these diseases. Experimental results show that the MobileNetV3-Small model (a variant of the MobileNet network) is the most optimal, offering fast training time, high accuracy, and acceptable levels of loss and error.

VQABG: Vietnamese question/answers benchmark generator for field-specific chatbot ground-truth dataset using EMINI (Exact Match wIth Numeric Information) indicator

Ngo-Ho Anh-Khoa, Vo Khuong-Duy, Ngo-Ho Anh-Khoi*

Abstract |

PDF

Currently, the application of generative Artificial Intelligence for developing specialized chatbots in Vietnamese is an inevitable trend. However, one of the most challenging aspects of assessing the quality of Vietnamese chatbot products is creating a specialized benchmark in a question-and-answer format. Typically, this benchmark is manually crafted by industry experts, which can be extremely costly. In contrast, for English, we can use bag-of-words model toolkits and grammatical structure architectures to generate appropriate questions automatically based on pre-existing answers from the original data. However, there is almost no complete model available for this task in Vietnamese. Regarding quality assessment, this is usually performed manually by experts using Human Evaluation (HE) indicators, which is also costly. Therefore, the aim of this study is to propose an algorithmic architecture specifically designed for the Vietnamese language. This architecture will automatically generate a set of question-and-answer queries to create a benchmark, as well as facilitate the development of a mechanism for automatic, straightforward, cost-effective, and accurate quality assessment for Vietnamese chatbots. We refer to this system as the Vietnamese Question/Answers Benchmark Generator (VQABG) and propose an innovative evaluation indicator called the Exact Match with Numeric Information (EMINI).

Product recommendation for online sales systems based on transaction sessions

Mai Le Bich Tuyen, Thanh Hai Nguyen, Tran Thanh Dien*

Abstract |

PDF

The advancement of information technology has influenced and accelerated the growth of many economic fields and industries, including e-commerce. This is one area of information technology that has rapidly developed and gained widespread popularity due to the significant benefits it offers to the community. Today, the need for online shopping is increasing because of its convenience and timesaving, especially for busy people. Historical session data plays an important role in helping sales systems recommend orders to customers that suit their personal preferences. This study proposes models for recommending products on online sales systems based on transaction sessions using memory-based collaborative filtering methods, including user-based and item-based, and model-based collaborative filtering methods including SVD and KNN. The experimental results show that the SVD model has a better rating prediction performance than other techniques. Therefore, it is probably proposed for product recommendations on online sales systems.

A multivariate analysis of the early dropout using classical machine learning and local interpretable model-agnostic explanations

Thanh Hai Nguyen*, Phuong Le, Tuyen Thanh Thi Nguyen, Anh Kim Su

Abstract |

PDF

Student dropout rates can have a significant negative impact on both the development of educational institutions and the personal growth of students. Consequently, many institutions are focused on identifying key factors that contribute to dropout and implementing strategies to mitigate them. This study aims to predict student dropout rates using classical machine learning algorithms while analyzing the key factors influencing these outcomes in higher education. The dataset includes demographic, socioeconomic, and academic information from various sources. Additionally, the study leverages the Local Interpretable Model-Agnostic Explanations (LIME) model to provide insights into the predictions, offering a clearer understanding of the factors driving dropout decisions. This knowledge is crucial for identifying influential factors and, more importantly, enhancing early intervention strategies and policies in educational settings, ultimately reducing dropout rates.

Exploring multiple optimization algorithms in transfer learning with EfficientNet models for agricultural insect classification

Hoang-Tu Vo*, Nhon Nguyen Thien, Kheo Chau Mui, Huan Lam Le, Phuc Pham Tien

Abstract |

PDF

Dangerous insects are a significant risk to the global agricultural industry, threatening food security, economic stability, and crop quality. This study investigates the impact of multiple optimization algorithms within transfer learning, employing EfficientNet models for the classification of agricultural insects. The explored optimization algorithms include Adam, Adamax, AdamW, RMSprop, and SGD, while utilizing the EfficientNetB0, EfficientNetB3, EfficientNetB5, and EfficientNetB7 architectures. Experimental results show notable performance differences between optimization algorithms across all EfficiencyNet models in the study. Among the measured metrics are precision, recall, f1-score, accuracy, and loss, the AdamW optimizer consistently demonstrates superior performance compared to other algorithms. The findings underscore the critical influence of optimization algorithms in enhancing classification accuracy and convergence within transfer learning scenarios. Additionally, the study employs various visualization techniques, such as Gradient-weighted Class Activation Mapping (Grad-CAM) to enhance the interpretation of the image classification model’s results. By focusing on these methodologies, this research aims to improve the model’s performance, optimize its capabilities, and ultimately contribute to effective pest management strategies in agriculture, safeguarding crop yields, farmer livelihoods, and global food security.

Fine-tuned PhoBERT for sentiment analysis of Vietnamese phone reviews

Ngo Minh Tan, Ngo Ba Hung*, Stuchilin Vladimir Valerievich

Abstract |

PDF

This paper presents an exploration of sentiment analysis applied to Vietnamese phone reviews, leveraging the PhoBERT model. While significant advancements have been made in sentiment analysis for English and other widely spoken languages, Vietnamese remains relatively under investigated. Our study addresses this gap by constructing a comprehensive dataset that integrates data from the UIT-ViSFD dataset and data collected through web scraping. We experimented with various models including naive Bayes, Support Vector Machine, and PhoBERT, utilizing multiple data preprocessing techniques. PhoBERT, a state-of-the-art pre-trained language model specifically designed for Vietnamese, demonstrated superior performance. The final PhoBERT model with optimized preprocessing achieved an accuracy of 92.74%, highlighting its efficacy in accurately identifying sentiments.

	All	Since 2019
Citations	1086	997
h-index	14	13
i10-index	20	17