• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.3
• /
• pp.133-140
• /
• 2022

We investigate the performance of deep learning-based Korean language models on a task of predicting the score range of Korean essays written by foreign students. We construct a data set containing a total of 304 essays, which include essays discussing the criteria for choosing a job ('job'), conditions of a happy life ('happ'), relationship between money and happiness ('econ'), and definition of success ('succ'). These essays were labeled according to four letter grades (A, B, C, and D), and a total of eleven essay score range prediction experiments were conducted (i.e., five for predicting the score range of 'job' essays, five for predicting the score range of 'happiness' essays, and one for predicting the score range of mixed topic essays). Three deep learning-based Korean language models, KoBERT, KcBERT, and KR-BERT, were fine-tuned using various training data. Moreover, two traditional probabilistic machine learning classifiers, naive Bayes and logistic regression, were also evaluated. Experiment results show that deep learning-based Korean language models performed better than the two traditional classifiers, with KR-BERT performing the best with 55.83% overall average prediction accuracy. A close second was KcBERT (55.77%) followed by KoBERT (54.91%). The performances of naive Bayes and logistic regression classifiers were 52.52% and 50.28% respectively. Due to the scarcity of training data and the imbalance in class distribution, the overall prediction performance was not high for all classifiers. Moreover, the classifiers' vocabulary did not explicitly capture the error features that were helpful in correctly grading the Korean essay. By overcoming these two limitations, we expect the score range prediction performance to improve.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.1309-1314
• /
• 2003

If an agent has a learning ability using previous knowledge, then it is expected that the agent can speed up learning by interacting with environment. In this paper, we present an improved reinforcement learning algorithm using domain knowledge which can be represented by problem-independent features and their classifiers. Here, neural networks are employed as knowledge classifiers. To show the validity of our proposed algorithm, computer simulations are illustrated, where navigation problem of a mobile robot and a micro aerial vehicle(MAV) are considered.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.539-541
• /
• 2023

COVID-19 is a viral pandemic disease that spreads widely all around the world. The only way to identify COVID-19 patients at an early stage is to stop the spread of the virus. Different approaches are used to diagnose, such as RT-PCR, Chest X-rays, and CT images. However, these are time-consuming and require a specialized lab. Therefore, there is a need to develop a time-efficient diagnosis method to detect COVID-19 patients. The proposed machine learning (ML) approach predicts the presence of coronavirus based on clinical symptoms. The clinical dataset is collected from the Israeli Ministry of Health. We used different ML classifiers (i.e., XGB, DT, RF, and NB) to diagnose COVID-19. Later, classifiers are optimized with the Bayesian hyperparameter optimization approach to improve the performance. The optimized RF outperformed the others and achieved an accuracy of 97.62% on the testing data that help the early diagnosis of COVID-19 patients.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.4
• /
• pp.860-880
• /
• 2024

Breast cancer is a major health concern affecting women and men globally. Early detection and accurate classification of breast cancer are vital for effective treatment and survival of patients. This study addresses the challenge of accurately classifying breast tumors using machine learning classifiers such as MLP, AdaBoostM1, logit Boost, Bayes Net, and the J48 decision tree. The research uses a dataset available publicly on GitHub to assess the classifiers' performance and differentiate between the occurrence and non-occurrence of breast cancer. The study compares the 10-fold and 5-fold cross-validation effectiveness, showing that 10-fold cross-validation provides superior results. Also, it examines the impact of varying split percentages, with a 66% split yielding the best performance. This shows the importance of selecting appropriate validation techniques for machine learning-based breast tumor classification. The results also indicate that the J48 decision tree method is the most accurate classifier, providing valuable insights for developing predictive models for cancer diagnosis and advancing computational medical research.

• International Journal of Computer Science & Network Security
• /
• v.21 no.9
• /
• pp.297-302
• /
• 2021

The development of various phishing websites enables hackers to access confidential personal or financial data, thus, decreasing the trust in e-business. This paper compared the detection techniques utilizing URL-based features. To analyze and compare the performance of supervised machine learning classifiers, the machine learning classifiers were trained by using more than 11,005 phishing and legitimate URLs. 30 features were extracted from the URLs to detect a phishing or legitimate URL. Logistic Regression, Random Forest, and CatBoost classifiers were then analyzed and their performances were evaluated. The results yielded that CatBoost was much better classifier than Random Forest and Logistic Regression with up to 96% of detection accuracy.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.3
• /
• pp.704-719
• /
• 2024

Botnet pandemics are becoming more prevalent with the growing use of mobile phone technologies. Mobile phone technologies provide a wide range of applications, including entertainment, commerce, education, and finance. In addition, botnet refers to the collection of compromised devices managed by a botmaster and engaging with each other via a command server to initiate an attack including phishing email, ad-click fraud, blockchain, and much more. As the number of botnet attacks rises, detecting harmful activities is becoming more challenging in handheld devices. Therefore, it is crucial to evaluate mobile botnet assaults to find the security vulnerabilities that occur through coordinated command servers causing major financial and ethical harm. For this purpose, we propose a hybrid analysis approach that integrates permissions and API and experiments on the machine-learning classifiers to detect mobile botnet applications. In this paper, the experiment employed benign, botnet, and malware applications for validation of the performance and accuracy of classifiers. The results conclude that a classifier model based on a simple decision tree obtained 99% accuracy with a low 0.003 false-positive rate than other machine learning classifiers for botnet applications detection. As an outcome of this paper, a hybrid approach enhances the accuracy of mobile botnet detection as compared to static and dynamic features when both are taken separately.

• ETRI Journal
• /
• v.43 no.1
• /
• pp.95-108
• /
• 2021

Application (app) ratings are feedback provided voluntarily by users and serve as important evaluation criteria for apps. However, these ratings can often be biased owing to insufficient or missing votes. Additionally, significant differences have been observed between numeric ratings and user reviews. This study aims to predict the numeric ratings of Google apps using machine learning classifiers. It exploits numeric app ratings provided by users as training data and returns authentic mobile app ratings by analyzing user reviews. An ensemble learning model is proposed for this purpose that considers term frequency/inverse document frequency (TF/IDF) features. Three TF/IDF features, including unigrams, bigrams, and trigrams, were used. The dataset was scraped from the Google Play store, extracting data from 14 different app categories. Biased and unbiased user ratings were discriminated using TextBlob analysis to formulate the ground truth, from which the classifier prediction accuracy was then evaluated. The results demonstrate the high potential for machine learning-based classifiers to predict authentic numeric ratings based on actual user reviews.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.7
• /
• pp.9-15
• /
• 2009

The fuzzy weighted mean classifier is one of the most common classification models and could achieve high performance by adjusting the weights. However, the weights were generally decided based on the experience of experts, which made the resulting classifiers to suffer the lack of consistency and objectivity. To resolve this problem, in this paper, a weight deciding method based on the statistics of the data is introduced, which ensures the learned classifiers to be consistent and objective. To investigate the effectiveness of the proposed methods, Iris data set available from UCI machine learning repository is used and promising results are obtained.

• Journal of Information Processing Systems
• /
• v.13 no.4
• /
• pp.928-940
• /
• 2017

State-of-the-art speaker recognition systems may work better for the English language. However, if the same system is used for recognizing those who speak different languages, the systems may yield a poor performance. In this work, the decisions of a Gaussian mixture model-universal background model (GMM-UBM) and a learning vector quantization (LVQ) are combined to improve the recognition performance of a multilingual speaker identification system. The difference between these classifiers is in their modeling techniques. The former one is based on probabilistic approach and the latter one is based on the fine-tuning of neurons. Since the approaches are different, each modeling technique identifies different sets of speakers for the same database set. Therefore, the decisions of the classifiers may be used to improve the performance. In this study, multitaper mel-frequency cepstral coefficients (MFCCs) are used as the features and the monolingual and cross-lingual speaker identification studies are conducted using NIST-2003 and our own database. The experimental results show that the combined system improves the performance by nearly 10% compared with that of the individual classifier.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.5 no.3
• /
• pp.263-267
• /
• 2005

Building classifiers for financial real-world classification problems is often plagued by severely overlapping and highly skewed class distribution. New performance measures such as receiver operating characteristic (ROC) curve and area under ROC curve (AUC) have been recently introduced in evaluating and building classifiers for those kind of problems. They are, however, in-effective to evaluation of classifier's discrimination performance in a particular class of the classification problems that interests lie in only a local operating range of the classifier, In this paper, a new method is proposed that enables us to directly improve classifier's discrimination performance at a desired local operating range by defining and optimizing a partial area under ROC curve or domain-specific curve, which is difficult to achieve with conventional classification accuracy based learning methods. The effectiveness of the proposed approach is demonstrated in terms of fraud detection capability in a real-world fraud detection problem compared with the MSE-based approach.