检索结果-内蒙古大学图书馆

A Comparison Study of classifier algorithms for Mobile-phone's Accelerometer Based Activity Recognition

Procedia Engineering 2012年 41卷 224-229页

作者： Media Anugerah Ayu Siti Aisyah Ismail Ahmad Faridi Abdul Matin Teddy Mantoro Intelligent Environment Research Group (INTEG Lab) Faculty of Information and Communication Technology International Islamic University Malaysia 53100 Kuala LumpurMalaysia Advanced Informatics School Semarak International Campus Universiti Teknologi Malaysia 53000 Kuala LumpurMalaysia

Nowadays, many mobile phones have been equipped with sensors to enable the delivery of advanced features/services to the users. Accelerometer is one of the sensors that embedded to several types of mobile phones. Our earlier research has shown that data from mobile-phone embedded accelerometer can be used for activity recognition purpose [1] . As a continuation of the research towards the search for a suitable and reliable algorithm for real-time activity recognition using mobile phone, an evaluation and comparison study of the performance of seven different categories of classifier algorithms in classifying user activities were conducted. Five basic human activities (jogging, jumping, sitting, standing, and walking) were tested. The training and testing data were done using Weka 3.6.6 data mining tool. The overall accuracy rate for classifier training managed to exceed 96% and exceeded 90% for classifier testing, which are very encouraging results.

关键词： Activity recognition Accelerometer classifier algorithms Activity-aware Accelerometer based mobile phone

来源：评论

学校读者我要写书评

暂无评论

AUTOMATED CLASSIFICATION OF ELECTRICAL APPLIANCES ON A PRODUCTION LINE

引用

NDT INTERNATIONAL 1988年第6期21卷 438-447页

作者： PELEG, K KOREM, A SHNEIDER, M TADIRAN CORP REHOVOT ISRAEL

User simulation inspection of every appliance on a production line is time consuming and expensive. A more effective way is to use sensors for fast indirect measurements of selected quality indexes, ie appliance features, which may be used for automatic on-line inspection and classification, by correlation to manual inspection. Feature vectors for classifying electrical appliances tend to form overlapping, irregular amorphous clusters in a multidimensional feature space. Three classifier algorithms were formulated to address this difficult classification problem, which is aggravated by the requirement that almost no bad units should be misclassified as good ones. The discriminatory power of two or all three classifiers is combined by a classifier voting strategy. The different classifiers and voting strategies are compared in terms of four performance indexes, predicting the percentage of bad units sent to the customer, percentage of good units rejected as bad ones, a cost-weighted class contamination index and the expected percentage of correctly classified units. Practical application is implemented on a feature data base of several hundred labelled refrigerators, whereby it is demonstrated that three classifier c voting will practically never misclassify a bad unit as a good one.

关键词： product inspection classifier algorithms classifier voting

来源：评论

学校读者我要写书评

暂无评论

Efficient Techniques for Attack Detection Using Different Features Selection algorithms and classifiers

引用

WIRELESS PERSONAL COMMUNICATIONS 2018年第4期100卷 1689-1706页

作者： Ghazy, Rania A. El-Rabaie, El-Sayed M. Dessouky, Moawad I. El-Fishawy, Nawal A. Abd El-Samie, Fathi E. El Sadat Univ Sadat Egypt Menoufia Univ Fac Elect Engn Dept Elect & Elect Commun Menoufia 32952 Egypt Menoufia Univ Fac Elect Engn Dept Comp Sci & Engn Menoufia 32952 Egypt

With the growth and benefits of network usage, securing the networks by using anomaly intrusion detection systems (IDS) against unknown intrusions has become an important issue. The first step of protecting any network is the detection of attacks. In this paper, we concentrate on four attacks;denial of service (DoS), probing, remote-to-local, and user-to-root attacks. We depend on features extracted from (NSL-KDD) dataset for these attacks. We investigate the performance of the attack detection process for several numbers of features using various subset-based feature selection techniques aiming to find the optimum collection of features for detecting each attack with an appropriate classifier. Simulation results reveal that redundant features can be eliminated from the attack detection process, and that we can determine the most useful set of features for a certain classifier, which enhances the IDS performance.

关键词： Network intrusion detection (NID) NSL-KDD dataset Features selection classifier algorithms Denial-of service (DoS) Probing (Prob) Remote-to-local (R2L) User-to-root (U2R)

来源：评论

学校读者我要写书评

暂无评论

A multimodal biomarker for concussion identification, prognosis and management

引用

COMPUTERS IN BIOLOGY AND MEDICINE 2018年 102卷 95-103页

作者： Jacquin, Arnaud Kanakia, Saloni Oberly, Doug Prichep, Leslie S. BrainScope Co Inc Algorithm Dev Bethesda MD USA BrainScope Co Inc Clin Affairs Bethesda MD USA NYU Sch Med Dept Psychiat New York NY USA

Background Prompt, accurate, objective assessment of concussion is crucial, particularly for children/adolescents and young adults. While there is currently no gold standard for the diagnosis of concussion, the importance of multidimensional/multimodal assessments has recently been emphasized. Methods: Concussed subjects (N = 177), matched controls (N = 187) and healthy volunteers (N = 204) represented a convenience sample of male and female subjects between the ages of 13 and 25 years, enrolled at 29 Colleges and 19 High Schools in the US. Subjects were tested at time of injury and at multiple time points during recovery. Assessments included EEG, neurocognitive tests and standard concussion assessment tools. Multimodal classifiers to maximally separate controls from concussed subjects with prolonged recovery (>= 14 days) were derived using quantitative EEG, neurocognitive and vestibular measures, informed feature reduction and a Genetic Algorithm methodology for classifier derivation. The methodology protected against overtraining using an internal cross-validation framework. An enhanced multimodal Brain Function Index (eBFI) was derived from the classifier output and mapped to a percentile scale which expressed the index relative to non-injured controls. Results: At time of injury eBFIs were significantly different between controls and concussed subjects with prolonged recovery, showing return to non-concussed levels at return-to-play plus 45 days. For the combined concussed population, and for the short recovery subjects, a more rapid recovery was seen. Conclusions: This multivariate, multimodal, objective index of brain function impairment can potentially be used, along with other tools, to aid in diagnosis, assessment, and tracking of recovery from concussion.

关键词： TBI Brain function index Enhanced BFI Concussion Sport-related concussion classifier algorithms Quantitative brain activity Electrophysiology of TBI Neurocognitive assessments Return to play

来源：评论

学校读者我要写书评

暂无评论

Classification algorithms for the identification of structural injury in TBI using brain electrical activity

引用

COMPUTERS IN BIOLOGY AND MEDICINE 2014年第0期53卷 125-133页

作者： Prichep, Leslie S. Dastidar, Samanwoy Ghosh Jacquin, Arnaud Koppes, William Miller, Jonathan Radman, Thomas O'Neil, Brian Naunheim, Rosanne Huff, J. Stephen NYU Sch Med Dept Psychiat Brain Res Labs New York NY USA BrainScope Co Inc Algorithm Dev Dept Bethesda MD USA Wayne State Univ Sch Med Dept Emergency Med Detroit MI USA Washington Univ Sch Med Div Emergency Med St Louis MO USA Univ Virginia Dept Emergency Med Charlottesville VA USA Univ Virginia Dept Neurol Charlottesville VA USA

Background: There is an urgent need for objective criteria adjunctive to standard clinical assessment of acute Traumatic Brain Injury (TBI). Details of the development of a quantitative index to identify structural brain injury based on brain electrical activity will be described. Methods: Acute closed head injured and normal patients (n=1470) were recruited from 16 US Emergency Departments and evaluated using brain electrical activity (EEG) recorded from forehead electrodes. Patients had high GCS (median=15), and most presented with low suspicion of brain injury. Patients were divided into a CT positive (CT+) group and a group with CT negative findings or where CT scans were not ordered according to standard assessment (CT-/CT_NR). Three different classifier methodologies, Ensemble Harmony, Least Absolute Shrinkage and Selection Operator (LASSO), and Genetic Algorithm (GA), were utilized. Results: Similar performance accuracy was obtained for all three methodologies with an average sensitivity/specificity of 97.5%/59.5%, area under the curves (AUC) of 0.90 and average Negative Predictive Validity (NPV) > 99%. Sensitivity was highest for CT+ cases with potentially life threatening hematomas, where two of three classifiers were 100%. Conclusion: Similar performance of these classifiers suggests that the optimal separation of the populations was obtained given the overlap of the underlying distributions of features of brain activity. High sensitivity to CT+ injuries (highest in hematomas) and specificity significantly higher than that obtained using ED guidelines for imaging, supports the enhanced clinical utility of this technology and suggests the potential role in the objective, rapid and more optimal triage of TB! patients. Published by Elsevier Ltd.

关键词： TBI Acute traumatic brain injury classifier algorithms Genetic algorithms Quantitative brain activity Electrophysiology of TBI TBI triage Structural brain injury CT plus TBI

来源：评论

学校读者我要写书评

暂无评论

Response to letter to the Editor regarding 'Classification algorithms for the identification of structural injury in TBI using brain electrical activity'

引用

COMPUTERS IN BIOLOGY AND MEDICINE 2015年 65卷 147-148页

作者： Prichep, Leslie S. Dastidar, Samanwoy Ghosh Jacquin, Arnaud Koppes, William Miller, Jonathan O'Neil, Brian Naunheim, Roseanne Huff, J. Stephen NYU Sch Med Dept Psychiat Brain Res Labs New York NY 10010 USA BrainScope Co Inc Algorithm Dev Bethesda MD USA Wayne State Univ Sch Med Dept Emergency Med Detroit MI USA Washington Univ Sch Med Div Emergency Med St Louis MO USA Univ Virginia Dept Emergency Med Charlottesville VA USA Univ Virginia Dept Neurol Charlottesville VA USA

来源：评论

学校读者我要写书评

暂无评论

Evaluation of Four classifiers as Cost Function for Indoor Location Systems

Evaluation of Four Classifiers as Cost Function for Indoor L...

引用

5th International Conference on Ambient Systems, Networks and Technologies (ANT) / 4th International Conference on Sustainable Energy Information Technology (SEIT)

作者： Galvan-Tejada, Carlos E. Garcia-Vazquez, Juan P. Garcia-Ceja, Enrique Carrasco-Jimenez, Jose C. Brena, Ramon F. Inst Tecnol Monterrey Monterrey Nuevo Leon Mexico

In our previous research work, we proposed a methodology that uses magnetic-field and multivariate methods to estimate user location in an indoor environment. In this paper, we propose the use of this methodology to evaluate the performance of four different classification algorithms: Random Forest, Nearest Centroid, K Nearest Neighbors and Artificial Neural Networks;each classifier will be considered as a cost function of a genetic algorithm (GA) used in the feature selection process task of the methodology. The motivation to evaluate the algorithms of classification was that several ILSs use a classification algorithm in order to estimate the location of the user, but the classifiers performance vary from application to application. In order to evaluate the performance of each classification algorithm, the following issues were considered: (1) the time of the training phase to obtain the final classification algorithm;(2) the number of features needed for getting the model;(3) the type of the features from the final model;and (4) the sensitivity and specificity of the model. Our results indicate that Nearest centroid is the classfier algorithm that is best suited to be implemented in an end-user application given the obtained results on the evaluated criteria for the indoor location system (ILS). (C) 2014 Published by Elsevier B.V.

关键词： Indoor Location classifier algorithms Nearest Centroid K Nearest Neighbors Artificial Neural Networks Multivariate Methods

来源：评论

学校读者我要写书评

暂无评论

Semantic Clone Detection Using Machine Learning 15

Semantic Clone Detection Using Machine Learning

引用

15th IEEE International Conference on Machine Learning and Applications (ICMLA)

作者： Sheneamer, Abdullah Kalita, Jugal Univ Colorado Colorado Springs CO 80918 USA

ISBN: (纸本)9781509061679

If two fragments of source code are identical to each other, they are called code clones. Code clones introduce difficulties in software maintenance and cause bug propagation. In this paper, we present a machine learning framework to automatically detect clones in software, which is able to detect Types-3 and the most complicated kind of clones, Type-4 clones. Previously used traditional features are often weak in detecting the semantic clones The novel aspects of our approach are the extraction of features from abstract syntax trees (AST) and program dependency graphs (PDG), representation of a pair of code fragments as a vector and the use of classification algorithms. The key benefit of this approach is that our approach can find both syntactic and semantic clones extremely well. Our evaluation indicates that using our new AST and PDG features is a viable methodology, since they improve detecting clones on the IJaDataset2.0.

关键词： Code clones Software clones classifier algorithms Abstract syntax trees (AST) Program dependence graphs (PDG)

来源：评论

学校读者我要写书评

暂无评论

An Embedded Application System for Data Collection of Atmospheric Pollutants with a Classification Approach 14th

An Embedded Application System for Data Collection of Atmosp...

引用

14th Mexican International Conference on Artificial Intelligence (MICAI)

作者： Solorzano-Alor, Eduardo Jose Argueelles-Cruz, Amadeo Isabel Cajero-Lazaro, Maria Sanchez-Meraz, Miguel IPN CIDETEC Ctr Innovac & Desarrollo Tecnol Comp Mexico City 07738 DF Mexico IPN CIC Mexico City 07738 DF Mexico UPIICSA Mexico City 07738 DF Mexico IPN ESIME Mexico City 07738 DF Mexico

This paper shows the application of an embedded system with a wireless sensor network to collect atmospheric pollutants data obtained from sensors placed into micro-climates;such dataset provides the information requi... 详细信息

ISBN: (纸本)9783319270609;9783319270593

关键词： Embedded system Waspmote classifier algorithms Atmospheric pollutants Micro-climates

来源：评论

学校读者我要写书评

暂无评论

INSURANCE CLAIM RISK SCORING WITH MACHINE LEARNING algorithms: A CASE STUDY ON DEVELOPING A PREDICTIVE SYSTEM FOR ASSESSING CORPORATE CUSTOMERS' CLAIM RISK

INSURANCE CLAIM RISK SCORING WITH MACHINE LEARNING ALGORITHM...

引用

作者： Tuomi-Nikula, Tino LAPPEENRANTA UNIVERSITY OF TECHNOLOGY

学位级别：硕士

Predictive learning algorithms offer tools to automate and improve insurance risk management. The aim of this thesis is to study classification algorithms in risk scoring applications and to evaluate them in the creation of insurance claim risk scores from customer risk survey data. Another goal for the thesis is to quantify weights for risk survey questions using machine learning methods. This case study consists of a critical literature review and a quantitative analysis where the risk scoring systems were developed and their performance on test data evaluated. Logistic regression, Support Vector Machines, Extreme Gradient Boosted Trees (XGBT) and a Feed Forward Neural Network were used to predict claim risk from customer risk survey data. The results of the study show that with the test data, XGBT was the most accurate, predicting the risk class of a customer with 73% accuracy. Nonetheless, it is worth mentioning that all tested algorithms were suitable for the task. The results also suggest that machine learning risk prediction enables analysis of individual risk factors, and that feature importance metrics can be used to quantify weights for risk survey questions.

关键词： risk scoring machine learning insurance risk classifier algorithms predictive analytics

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：