Feature selection is a fundamental problem in machine learning and data mining. How to choose the most problem-related features from a set of collected features is essential. In this paper, a novel method using correlation coefficient clustering in removing similar/redundant features is proposed. The collected features are grouped into clusters by measuring their correlation coefficient values. The most class-dependent feature in each cluster is retained while others in the same cluster are removed. Thus, the most class-related and mutually unrelated features are identified. The proposed method was applied to two datasets: the disordered protein dataset and the Arrhythmia (ARR) dataset. The experimental results show that the method is superior to other feature selection methods in speed and/or accuracy. Detail discussions are given in the paper.
This research aimed at building an intelligent system that can detect abnormal behavior for the elderly at home. Active RFID tags can be deployed at home to help collect daily movement data of the elderly who carries an RFID reader. When the reader detects the signals from the tags, RSSI values that represent signal strength are obtained. The RSSI values are reversely related to the distance between the tags and the reader and they are recorded following the movement of the user. The movement patterns, not the exact locations, of the user are the major concern. With the movement data (RSSI values), the clustering technique is then used to build a personalized model of normal behavior. After the model is built, any incoming datum outside the model can be viewed as abnormal and an alarm can be raised by the system. In this paper, we present the system architecture for RFID data collection and preprocessing, clustering for anomaly detection, and experimental results. The results show that this novel approach is promising.
With the increasing number of emergencies, the crowd simulation technology has attracted wide attention in the recent years. Existing emergencies have shown that individuals are easy to be influenced by others' emotion during the evacuation. This will make it easier for people to aggregate together and increase security risks. Some of the existing evacuation models without considering emotion are therefore not suitable for describing crowd behaviors in emergencies. We propose a perception-based emotion contagion model and use multiagent technology to simulate crowd behaviors. Navigation points are introduced to guide the movement of the agents. Based on the proposed model, a prototype simulation system for crowd emotion contagion is developed. The comparative simulation experiments verify that the model can effectively deduct the evacuation time and crowd emotion contagion. The proposed model could be an assistant analysis method for crowd management in emergencies.Public places have been more crowded with the progress of global urbanization. Meanwhile, public safety and crisis management are facing unprecedented challenges with the increasing number of emergencies. In emergencies, overcrowded places become dangerous and would result in a large number of casualties. It has important and practical significance to simulate crowd behaviors in emergencies.After analyzing real incidents of emergencies, we find that existing emergency plans always have problems to cope with changeable emergencies. In emergencies, people easily generate negative emotions and will spread their panic to others. The panic emotion will lead to irrational behaviors. Exploring the law of crowd evacuation behaviors will help us develop a scientific crowd management plan and reduce the emergent evacuation risk. However, it is not suitable to observe crowd behaviors from evacuation drills, as those drills will be constrained by space, cost, or time and cannot be Comput Anim Virtual Worlds. 2018;29:e1817.wileyonlinelibrary.com/journal/cav
With the rapid development of distance learning and the XML technology, metadata play an important role in e-Learning. Nowadays, many distance learning standards, such as SCORM, AICC CMI, IEEE LTSC LOM and IMS, use metadata to tag learning materials. However, most metadata models are used to define learning materials and test problems. Few metadata models are dedicated to assessment. In this paper, the authors propose an assessment metadata model for e-Learning operations. With support from assessment metadata, we can incorporate measured aspects of the following list into the metadata description at the question cognition level, the item difficulty index, the item discrimination index, the questionnaire style and the question style. The assessment analysis model provides analytical suggestions for individual questions, summary of test results and cognition analysis. Analytical suggestions provide teachers information about why a question is not appropriate. Summary of test results improves the teacher's view of student learning status immediately. Items missing from the teaching materials can be identified by cognition analysis. In this research, the authors propose an enhanced metadata model and an implemented system based on our model. With metadata support, metadata can help teachers in authoring examination.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.