Research and Development on Information and Communication Technology

Một phương pháp phân cụm bán giám sát mờ đồng huấn luyện trên dữ liệu đa khung nhìn

Tue, 26 Sep 2023 00:00:00 +0700

In today’s practical reality, multi-view data is increasingly prevalent. Multi-view data refers to a type of data that encompasses multiple perspectives or viewpoints of an object. Data within each individual view possesses specific attributes dedicated to knowledge discovery and provides information on the same subject with varying degrees of accuracy and reliability. Combining various types of information from different views can yield a more comprehensive and accurate representation of objects, thereby improving data analysis and decision-making. Multi-view clustering has emerged as a research direction that has garnered the interest of scientists in recent years. However, there has been no research focusing on semi-supervised fuzzy clustering combined with co-training algorithms to assess the accuracy and quality of clustering on multi-view datasets. This paper proposes a novel method in semi-supervised clustering, utilizing co-training algorithms on multi-view data collected from a data source. Additionally, the paper provides experimental results to evaluate the effectiveness and accuracy of the proposed algorithm.

Về một thuật toán gia tăng tìm tập rút gọn trên bảng quyết định khi loại bỏ tập đối tượng

Tue, 26 Sep 2023 00:00:00 +0700

Feature selection or attribute reduction for decision information systems has long been considered a key and indispensable problem in data mining and analysis. Some approaches based on rough set theory and extensions have brought many attribute reduction methods with impressive efficiency. However, up to now, some attribute reduction methods according to intuitionistic fuzzy sets have not received much interest. The advance of this approach is the ability to improve classification performance on noisy and inconsistent decision tables. This paper starts from a distance measure between two intuitionistic fuzzy partitions and then proposes an effective attribute reduction algorithm. Specifically, we first design an attribute reduction algorithm on the decision table without change. Next, we construct an incremental algorithm to process the decision table when deleting an object set. Some experimental results have shown that our proposed methods have superior performance to methods based on the rough set and fuzzy set in terms of the size of the reduct and the classification efficiency.

Thuật toán song song khai thác itemset lợi nhuận phổ biến Skyline

Nguyễn Mạnh Hùng, Thi Thuy Tram Nguyen — Tue, 26 Sep 2023 00:00:00 +0700

Skyline common-utility element sets (SFUIs) can provide more useful information for decision-making by considering both their frequency and their benefits. Since the Skyline utility-common element set mining problem was proposed by Goyal V. and colleagues in 2015, up to now, many sequential algorithms have been proposed to improve mining performance. However, most algorithms have poor performance when exploiting today’s popular large data sets. In this paper, we propose a parallel algorithm called ParaSFUI-UF based on the sequential algorithm SFUI-UF, which is the most effective algorithm for exploiting the Skyline common-benefit element set today. Experimental results show that
the ParaSFUI-UF algorithm outperforms the SFUI-UF algorithm.

Nghiên cứu phương pháp phát hiện va chạm của cánh tay robot cộng tác 6 bậc tự do

Hung Nghiem Van, Van Can Nguyen — Thu, 05 Oct 2023 00:00:00 +0700

Cobots are robots that can directly contact humans, simultaneously promoting the advantages of both humans and robots to increase work efficiency. Cobots operate friendly and interactive with humans because they are programmed to detect collisions safely. Therefore, the need to accurately and quickly detect collisions of cobot arms is a topic that attracts the attention of many researchers. The our proposed technique using the SVMR model and the 1D CNN model is tested and gives good collision detection results with CURA6 cobot arm on the Intema’s dataset.

Thuật toán song song khai thác nhanh các mẫu trọng số hữu ích phổ biến từ cơ sở dữ liệu định lượng động

Nguyen Le, Ham Nguyen, Minh Nguyen — Thu, 12 Oct 2023 00:00:00 +0700

In recent years, the problem of mining frequent weighted utility patterns has been receiving research attention. This is a variation of the pattern mining problem. Many effective methods have been proposed to solve this problem. However, when the data is extensive and the weights of items change frequently, the algorithms take much time during the mining process. If we take advantage of the parallel computing capabilities of computing systems or multi-core processors, we can improve the mining time of algorithms. This paper presents a parallel solution operating on multi-core processors, named pdFWUNL, to exploit frequent weighted utility patterns from dynamic quantitative databases with variable item
weight. The experimental results show that the mining time of our parallel algorithm, pdFWUNL, is more efficient than the best available sequential method.

Social Network Recommendations for Friends with Neo4j Graph Database

Thuy Pham Thi Thu, Thanh Nguyen Thi Thai, Hwa Soo Kim — Fri, 13 Oct 2023 00:00:00 +0700

In recent years, along with the development of the internet, the number of social network users is increasing day by day. Through social networks, users can share, exchange information or make friends with each other. However, with new relationships, users often have a need to assess the credibility of a new friend before making friends on social networks. This paper proposes a friend recommendation system on social networks based on how to calculate the reliability between users. The recommendation system is implemented using the Neo4j graph database. The results of the truth algorithm proposed in this paper are higher than other similar algorithms.

Một thuật toán định tuyến cân bằng năng lượng trong mạng cảm biến không dây dựa trên SDN

Huy Lê Đức, Binh Le Huu, Công Đỗ Thành, Giang Nguyễn Đỗ Hoàng — Tue, 28 Nov 2023 10:35:25 +0700

In wireless sensor networks (WSN), it is necessary to use energy efficiently to extend the operating time of sensor nodes. In this study, we propose a routing algorithm that considers the energy consumption between sensor nodes. The goal of the proposed algorithm is to balance energy consumption and minimize the number of nodes that must consume a large amount of energy to increase their uptime. Our method constructs a weight function for wireless links that contains parameters for the energy remaining at the nodes. Then, a centralized routing mechanism based on a software-defined network (SDN) architecture is used to find the best weight route for data transfer. Simulation results on OMNeT++ show that the proposed algorithm increases the uptime of nodes and network throughput compared to the current popular routing algorithms.