UniMHe: Unified Multi Hyperedge Prediction A Case Study on Crime Dataset DOI

Melike Yildiz Aktas,

Lulwah Alkulaib, Chang‐Tien Lu

и другие.

2021 IEEE International Conference on Big Data (Big Data), Год журнала: 2023, Номер unknown, С. 5134 - 5139

Опубликована: Дек. 15, 2023

Edge prediction is a fundamental challenge in network science, with broad applications, notably social networks. It plays crucial role unveiling complex system dynamics by forecasting connections between entities. Our paper introduces UniMHe (Unified Multi Hyperedge Prediction), novel framework for predicting multiple hyperedges associated each node using hypergraph representations. We present case study focused on crime analysis, where reveals intricate patterns criminal activities, including types, locations, and seasonal variations. research leverages extensive historical data encompassing geographical information, timestamps, points of interest, categories. In an evaluation, we benchmark against state-of-the-art deep learning techniques, highlighting its superior performance. These findings underscore the significance across various domains problem-solving scenarios.

Язык: Английский

HDLCP: Experimental Analysis and Development of Hybrid Deep Learning Methodology for Crime Scenario Assesment and Prediction DOI

M. Tamilselvi,

C.N. Ravi,

S. Jayasudha

и другие.

Опубликована: Март 1, 2024

For a variety of reasons, prediction is utilized in nearly every industry. Many societal functions, like as crime prediction, are served by it. Data mining tools abound when it comes to traditional prediction. These approaches lack accuracy dealing with new kinds data and somewhat outdated. They take lot time well. In place the antiquated methods, Artificial Neural Networks function This study employs Hybrid Deep Learning based Crime Prediction (HDLCP) model forecast criminal activity; then tested for using Decision Tree (DT) technique cross-validation. Using current datasets, additional information anticipated be extracted. Criminal activity perilous pervasive, affecting societies all across globe. Life expectancy, GDP growth, national prestige impacted rates. New methods sophisticated technologies required enhance analytics order safeguard communities ensure safety society whole. A number probabilities certain area may studied, detected, predicted suggested approach. approaches, author explains many forms

Язык: Английский

Процитировано

8
Machine Learning Applied to Gender Violence: A Systematic Mapping Study DOI Creative Commons
Cristian-Camilo Pinto-Muñoz, Jhon-Alex Zuñiga-Samboni, Hugo Ordóñez

и другие.

Revista Facultad de Ingeniería, Год журнала: 2023, Номер 32(64), С. e15944 - e15944

Опубликована: Июнь 20, 2023

Machine Learning (ML) has positioned itself as one of the best tools to address different problems thanks its data processing capabilities, well models, algorithms, and predictive factors that help solve defined problems. Therefore, this article presents a systematic mapping from 2018 2023 focused on application ML gender-based violence. The methodology followed for study is based definition elements such research questions, search strings, bibliographic sources, inclusion exclusion criteria. results allow us understand benefits challenges using artificial intelligence, precisely branches, ML, combat in areas society, education, health, violence, among others. It also identifies countries where being researched contexts it applied to. discusses After conducting literature review, beneficial were found intelligence ML. obtained articles showed capacity improvements compared currently used systems. However, despite positive results, no evidence development an model or algorithm violence Colombia was review.

Язык: Английский

Процитировано

2
Construction And Performance Evaluation of Big Data Prediction Model Based on Fuzzy Clustering Algorithm in Cloud Computing Environment DOI Creative Commons

Et al. Yanhua Hu

Deleted Journal, Год журнала: 2024, Номер 19(4), С. 01 - 13

Опубликована: Янв. 25, 2024

In the evolving landscape of biomedical biometrics, where multimodal approaches are increasingly crucial for reliable user authentication, this research presents a comprehensive study. The primary focus is on construction and performance evaluation robust big data prediction model within cloud computing environment. advent has revolutionized field offering immense potential advanced analysis prediction. This development biometric in applications. proposed incorporation Reliable Discrete Variable Topology (RDVT) into model. RDVT introduces topological structure that enhances reliability ensures integrity information. training meticulously detailed, encompassing preprocessing, feature extraction, clustering, classification, evaluation. Additionally, integration fuzzy clustering algorithm model's ability to handle uncertainty imprecision data. advancement biometrics by introducing based rigorously assessed through extensive experimentation, including accuracy, precision, recall, F1-score measurements.

Язык: Английский

Процитировано

0
Optimization Algorithm of Intelligent Warehouse Management System Based on Reinforcement Learning DOI Creative Commons

Jianjun Zhou Jianjun Zhou

Deleted Journal, Год журнала: 2024, Номер 20(1), С. 219 - 231

Опубликована: Янв. 25, 2024

An Intelligent Warehouse Management System (IWMS) represents a technological leap forward in the realm of logistics and supply chain management. This sophisticated system integrates suite cutting-edge technologies, including artificial intelligence, machine learning, Internet Things, to revolutionize way warehouses operate. The primary focus is on construction performance evaluation robust big data prediction model within cloud computing environment. advent has revolutionized field Logistics, offering immense potential for advanced analysis prediction. research presents development IWMS Logistics applications. proposed incorporation Reliable Discrete Variable Topology (RDVT) into model. RDVT introduces topological structure that enhances reliability ensures integrity information. training are meticulously detailed, encompassing preprocessing, feature extraction, clustering, classification, evaluation. Additionally, integration fuzzy clustering with reinforcement learning algorithm model's ability handle uncertainty imprecision management data. advancement based rigorously assessed through extensive experimentation, accuracy, precision, recall, F1-score measurements.

Язык: Английский

Процитировано

0
Analysis of the Research Overview and Frontier Trend of Competitive Technical Intelligence Research Abroad DOI

思彤 刘

Statistics and Applications, Год журнала: 2023, Номер 12(05), С. 1283 - 1290

Опубликована: Янв. 1, 2023

Язык: Английский

Процитировано

0
UniMHe: Unified Multi Hyperedge Prediction A Case Study on Crime Dataset DOI

Melike Yildiz Aktas,

Lulwah Alkulaib, Chang‐Tien Lu

и другие.

2021 IEEE International Conference on Big Data (Big Data), Год журнала: 2023, Номер unknown, С. 5134 - 5139

Опубликована: Дек. 15, 2023

Edge prediction is a fundamental challenge in network science, with broad applications, notably social networks. It plays crucial role unveiling complex system dynamics by forecasting connections between entities. Our paper introduces UniMHe (Unified Multi Hyperedge Prediction), novel framework for predicting multiple hyperedges associated each node using hypergraph representations. We present case study focused on crime analysis, where reveals intricate patterns criminal activities, including types, locations, and seasonal variations. research leverages extensive historical data encompassing geographical information, timestamps, points of interest, categories. In an evaluation, we benchmark against state-of-the-art deep learning techniques, highlighting its superior performance. These findings underscore the significance across various domains problem-solving scenarios.

Язык: Английский

Процитировано

0