Coastal wetlands play a crucial role in the global carbon cycle, yet accurately estimating CO2 emissions from these ecosystems remains challenging. This paper presents a novel approach to estimate CO2 sequestration potential in coastal wetlands and mangrove forests, integrating remote sensing band analysis with Random Forest (RF) modeling and hyperparameter tuning through Grid Search, and Bayesian optimization. Utilizing multi-temporal Landsat 8 and Sentinel-2 imagery, we developed a linear regression (LR) model to predict CO2 concentrations, emphasizing the role of mangroves in mitigating greenhouse gases. Our RF model achieved an RMSE of 5.10e-12 and an R2 of 1.0, demonstrating exceptional accuracy. In contrast, Bayesian optimization yielded an RMSE of 64.49 with an R2 of 0.997, while Grid Search produced an RMSE of 63.29 and an R2 of 0.997. Additionally, our analysis revealed significant clustering of high-emission areas, primarily linked to degraded mangrove forests and aquaculture zones. These findings underscore the critical role of mangrove ecosystems in carbon sequestration and provide valuable insights for more effective climate change mitigation strategies in coastal wetlands.

Recommendation systems (RS systems) has been popular among research communities as they are contributing significantly to the e-commerce domain. This kind of systems helps entrepreneurs to introduce their products to prospective customers and they also help online shoppers find the products that might somehow satisfy their needs and preferences. Many different techniques and RS models were proposed with the intention to introduce products that mostly satisfy user’s preferences. In other words, the models that can improve performances of recommendation systems are significant. In this article, semantic similarity was employed in addition to Cosine similarity to measure the similarity between two users in RS models. K-Nearest Neighbor (KNN) was utilized to determine the most similar users, which are important to retrieve the most satisfied products. The comparison between RS models that employed semantic similarity and the RS model that does not take semantic similarity factor into consideration was conducted over the benchmark datasets MovieLens 100K. For system evaluation, Mean Absolute Error (MAE), Root Means Square Error (RMSE) and the Cohen’s kappa (KAPPA) metrics were used to evaluate the RS models. The comparisons between two groups of models including RS systems with regularized information and RS systems without regularized information were conducted. The experiment results show that the RS model that employed Word Embedding method as regularized information provided higher performances than the other RS systems.

In this study, we propose a novel approach analyzing the results of Hyperparameter Optimization (HPO) for Random Forest (RF) models by applying Energy Distance (ED), a metric based on pairwise Euclidean distances. This method provides a quantitative measure the similarities and differences between the configurations hyperparameters and their corresponding performance metrics. We use a dataset from a hyperparameter optimization experiment for RF, where we explore relationship between hyperparameter settings and model accuracy. The results indicate that Energy Distance can offer useful insights into the proximity of different hyperparameter configurations and help identify clusters similar configurations, which can be useful for model selection and optimization.

Machine learning models have been widely used in many applications in almost all areas of social life. Random forest is a supervised machine learning model that combines the results of multiple decision trees to achieve a single result using closure. Due to the ease of use and flexibility of the random forest machine learning model, there has been a push for its adoption in practical applications of both regression and classification problems. To fit the random forest machine learning model to different problems, the model parameters must be adjusted. Choosing the best parameter configuration for the model has a direct impact on the model’s performance. In this article, the parameters of the random forest model and parameter optimization algorithms are studied in detail. Furthermore, the study also tested different benchmark datasets to compare the performance of random forest model parameter optimization methods.

Phishing attacks, which exploit deceptive URLs to steal sensitive information, pose a significant cybersecurity threat. This paper presents a novel framework for phishing URL detection utilizing Convolutional Neural Networks (CNNs) optimized through advanced hyperparameter tuning. The proposed method involves preprocessing URLs by tokenizing them at the character level and padding sequences to ensure uniform input dimensions. A CNN model is constructed with embedding layers, convolutional operations, max pooling, and dense layers. Hyperparameter optimization is performed using the Hyperband algorithm from Kera's Tuner, systematically exploring various configurations including filter counts, kernel sizes, pooling strategies, and learning rates. Evaluated on a large dataset of phishing URLs, the model achieved a test accuracy of 80.50%, demonstrating the efficacy of CNNs for phishing detection and the benefits of thorough hyperparameter tuning. This study contributes to the field of cybersecurity by offering a robust, data-driven approach to phishing detection and providing insights into optimizing deep learning models for practical applications.

Numerous research studies have emphasized the significance of contextual information when it comes to recommender models.This importance is especially evident in the realm of e-commerce platforms, where recommender systems have been effectively suggesting products and services to users by integrating contextual data into their models.By doing so, these systems can better understand user preferences and behaviors during transactions on the platform.As a result, a growing number of platforms are now collecting evaluation values for products and services based on various user contexts, leading to the emergence of multi-context-based rating datasets.This presents a valuable opportunity to implement multi-criteria collaborative filtering models, which we propose as a solution.Our approach involves integrating user contextual rating data and conducting experiments using two sets of contextual evaluation datasets: De Paul Movie and In Car Music.The results demonstrate that the multi-criteria collaborative filtering model outperforms the single-contextbased collaborative filtering model in terms of accuracy.This study opens up promising avenues for future research aimed at further enhancing recommendation accuracy for customers on online sales platforms.

This paper proposes a new approach to solve the problem of lack of information in rating data due to new users or new items, or there is too little rating data of the user for items of the collaborative filtering recommendation models (CFR models). In this approach, we consider the similarity between users or items based on the lasso regression to build the CFR models. In the commonly used CFR models, the recommendation results are built only based on the feedback matrix of users. The results of our model are predicted based on two similarity calculated values: (1) the similarity calculated value based on the rating matrix; (2) the similarity calculated value based on the prediction results of the Lasso regression. The experimental results of the proposed models on two popular datasets have been processed and integrated into the recommenderlab package showed that the suggested models have higher accuracy than the commonly used CFR models. This result confirms that Lasso regression helps to deal with the lack of information in the rating data problem of the CFR models.

In recent studies on recommender models, association rules have been applied in many studies to improve the effectiveness of recommender models. However, these studies also reveal some drawbacks, such as the models take a considerable amount of time to generate association rules for large datasets; generation algorithms can ignore rules with the significant implication that affect the quality of recommender models. This study proposes collaborative filtering recommender models (CF models) based on association rules following an asymmetric approach of the statistical implicative analysis method to enhance the precision of recommender models. Through experiments on standard datasets and quality comparison with other CF models, we conclude that the proposed models based on the asymmetric relationship achieve better accuracy on the experimental datasets.

This article proposes a new method to overcome the sparse data problem of the collaborative filtering models (CF models) by considering the homologous relationship between users or items calculated on contextual attributes when we build the CF models. In the traditional CF models, the results are built only based on data from the user's ratings for items. The results of the proposed models are calculated on two factors: (1) the similar factors based on rating values; (2) the similar factors based on contextual attributes. The findings from the experimentation on two datasets DePaulMovie and InCarMusic, show that the proposed models have higher accuracy than the traditional CF models.

Context-aware recommender systems researches now concentrate on adjusting recommendation results for situations specific context of the users. These studies suggest many ways to integrate user contextual information into the recommendation process such as using topic hierarchies with matrix factorization techniques to improve context-aware recommender systems, measuring frequency-based similarity for context-aware recommender systems, collecting data from social networking to support context-aware recommender systems, and so on. However, these studies mainly focus on the development of context-aware recommendation algorithms to propose items to users in a particular situation and do not care about the extent of contextual involvement in the recommendation process to make recommendation results. In this article, we propose a new approach for context-aware recommender systems based on objective interestingness measures to consider the contextual relationship of the users in the recommendation process. Based on the experimental results on two standard datasets, the proposed model is more accurate than the traditional models.

In recent research, many approaches based on association rules have been proposed to improve the accuracy of recommender systems. These approaches are primarily based on Apriori data mining algorithm in order to generate the association rules and apply them to improving the recommendation results. However, these approaches also reveal some disadvantages of the system, such as taking a longer time for generating association rules; applying the Apriori algorithm on rating sparse matrix resulting in irrelevant information and causing poor recommendation results to target users and association rules generated primarily relying on given threshold of Support and Confidence measures leading to the focus on the majority of rules and ignoring the astonishment of rules to affect the recommendation results. In this study, we propose a new model for collaborative filtering recommender systems: The collaborative recommendation is based on statistical implication rules (IIR); Different from collaborative recommendation based on association rules (AR), the IIR predicts the items for users based on statistical implication rules generated from rating matrix and Implication intensity measures measuring the surprisingness of rules. To evaluate the effectiveness of the model, the experimental section is implemented on three real datasets to compare the results with some different effective models. The results show that the IIR has higher precision on the experimental datasets.

From the first appearance, recommender system in general and User-based collaborative filtering recommendation system have been developed greatly in technology and their application in life. In particular, recommender systems are used by many managers as an effective tool in order to support the business in various fields such as Amazon, Netflix and Pandora. However, the present generation of recommender systems has not fully met the requirements of users yet. In this paper, we propose a new approach for User-based collaborative filtering recommender system. The User-based collaborative filtering recommender system based on Implication statistic measures. In the system, we build a new similarity measures for two users are based on the Implication intensity measures. It is called statistical implicative similarity measures. Through experiments on two datasets MovieLense and MSWeb show that our similarity measures has fairly good results on User-based collaborative filtering model compared with traditional similarity measures as Pearson correlation, Cosine similarity, and Jaccard.

Hệ tư vấn được xem là một giải pháp hiệu quả có thể ứng phó với vấn đề bùng nổ thông tin do sự phát triển quá nhanh của các dịch vụ Internet và được ứng dụng rộng rãi trong nhiều lĩnh vực. Tuy nhiên, để thiết kế một hệ tư vấn có thể đáp ứng được nhu cầu của người dùng thì việc lựa chọn mô hình phù hợp cho hệ thống tư vấn và lựa chọn các giá trị tham số thích hợp cho mô hình luôn là một thách thức lớn của người thiết kế. Trong nghiên cứu này, chúng tôi đề xuất giải pháp lựa chọn mô hình và các giá trị tham số phù hợp cho bài toán tư vấn lọc cộng tác cụ thể. Để đánh giá các giải pháp đề xuất, chúng tôi tiến hành thực nghiệm trên ba tập dữ liệu chuẩn gồm: MovieLens, MSWeb và Jester5k. Kết quả thực nghiệm cho thấy các giải pháp của chúng tôi đề xuất có thể hỗ trợ nhà thiết kế, nhà nghiên cứu xác định được mô hình cũng như các giá trị tham số của mô hình cho bài toán tư vấn cụ thể của họ một cách nhanh chóng.

The objective interestingness measures play an important role in data mining because they are used for mining, filtering and ranking the patterns. However, there is no research that collects the measures fully as well as there is no tool that can: automatically calculate the interestingness values of the patterns by using those measures, and is the framework for rapidly developing the applications related to objective interestingness measures. This paper describes Interestingnesslab - a tool of the objective interestingness measures is developed in the R language. The main functions of the tool are: mining a set of association rules and presenting them by the cardinalities ( n,n_X,n_Y,n_XY ), calculating the interestingness value of an association rule according to 1 of 109 collected measures; calculating the interestingness values of the whole rule set in many measures selected by the user; discovering the tendencies in a data set and recommending the top N items to the user; and studying the specific behavior of a set of interestingness measures in the context of a specific dataset and in an exploratory data analysis perspective. With Interestingnesslab, the user can easily and quickly reuse its functions to develop his/her own applications.

The creation of the interestingness measures for evaluating the quality of the association rule - based knowledge plays an important role in the post-processing of the Knowledge Discovery from Databases. More and more interestingness measures are proposed by two approaches (subjective assessment and objective assessment), studying the properties or the attributes of the interestingness measures is important in understanding the nature of the objective interestingness measures. In this paper, we focus primarily on the objective interestingness measures to obtain a general view of recent researches on the nature of the objective interestingness measures, as well as complete a new classification on 109 selected objective interestingness measures on 6 criterions (independence, equilibrium, symmetry, variation, description, and statistics).

In recent years, the research cluster of objective interestingness measures has rapidly developed in order to assist users to choose the appropriate measure for their application. Researchers in this field mainly focus on three main directions: clustering based on the properties of the measures, clustering based on the behavior of measures and clustering tendency of variation in statistical implications. In this paper we propose a new approach to cluster the objective interestingness measures based on tendency of variation in statistical implications. In this proposal, we built the statistical implication data of 31 objective interestingness measures based on the examination of the partial derivatives on four parameters. From this data, two distance matrices of interestingness measures are established based on Euclidean and Manhattan distance. The similarity trees are built based on distance matrix that gave results of 31 measures clustering with two different clustering thresholds.