检索结果-南通市图书馆

A New Classifier for Imbalanced Data Based on a Generalized Density Ratio Model

在线全文

学校读者我要写书评

暂无评论

Communications in Mathematics and Statistics 2023年第2期11卷 369-401页

作者： Junjun Li Wenquan Cui Department of Statistics and Finance School of ManagementUniversity of Science and Technology of ChinaHefei 230026People’s Republic of China

Achieving higher true positive rate when decreasing false positive rate is always a great challenge to the imbalance learning community.This work combines penalized empirical likelihood method,lower bound algorithm and Nyströmmethod and applies these techniques along with kernel method to density ratio model.The resulting classifier,density ratio classifier(DRC),is a combination of kernelization,regularization,efficient implementation and threshold moving,all of which are critical to enable DRC to be an effective and powerful method for solving difficult imbalance problems.Compared with other methods,DRC is competitive in that it is widely applicable and it is simple and easy to use without additional imbalance handling skills.In addition,the convergence rate of the estimate of log density ratio is discussed as well.And the results of numerical analysis also show that DRC outperforms other methods in AUC and G-mean score.

关键词： Classifier Density ratio model Imbalance problems Kernel method roc curve

A Novel Industrial Intrusion Detection Method based on Threshold-optimized CNN-BiLSTM-Attention using roc curve

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

A Novel Industrial Intrusion Detection Method based on Thres...

Coronary bifurcation angle from 3-D predicts clinical outcomes after stenting bifurcation lesions

第三十九届中国控制会议

作者： Mindi Lan Jun Luo Senchun Chai Ruiqi Chai Chen Zhang Baihai Zhang School of Automation Beijing Institute of Technology School of Aerospace Transport and ManufacturingCranfield University

In recent years, many researchers have proposed many intrusion detection methods to protect the industrial network.However, there are two existing problems among them: one is that they only consider the overall accuracy rate（AC） while ignoring the problem of class imbalance;another one is that they have considered the problem of class imbalance, but the detection rate（DR） is low and false positive rate（FR） is high for minority classes. In order to improve AC and DR of minority classes,we propose a method called threshold-optimized CNN-BiLSTM-Attention that combines CNN-BiLSTM-Attention model, with threshold modification method based on receiver operating characteristic（roc） curve. In this method, we use CNN-BiLSTMAttention model as a classifier and modify threshold of the classifier through roc curve. To evaluate the proposed method, we have performed experiments on the standard industrial data set. And the experimental results show that the proposed method can improve AC and the DR of minority classes at low FR, which is better than other intrusion detection methods.

关键词： Industrial intrusion detection Class imbalance CNN-BiLSTM-Attention Threshold modification roc curve

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

Chinese Medical Journal 2012年第12期125卷 2083-2088页

作者： CHEN Shao-liang Teguh Santoso ZHANG Jun-jie YE Fei XU Ya-wei FU Qiang KAN Jing Chitprapai Paiboon ZHOU Yong DING Shi-qing Tak W Kwan Department of Cardiology Nanjing First Hospital NanjingMedical University Nanjing Jiangsu 210006 China Department of Cardiology Medistra Hospital University ofIndonesia Medical School Jakarta Indonesia Department of Cardiology Shanghai 10th People＇s HospitalShanghai 200072 China Department of Cardiology Xuzhou Central Hospital XuzhouJiangsu 221009 China Department of Cardiology Bangkok General Hospital BangkokThailand Department of Cardiology Zhangjiagang People＇s HospitalZhangjiagang Jiangsu 215600 China Department of Cardiology Huainan Xinhua Hospital HuainanAnhui 232038 China Department of Cardiology Beth Israel Hospital NewYork USA

Background The predictive value of bifurcation angle （BA） for worse events after stenting bifurcation lesions remains to be unknown. The present study was to investigate the dynamic change of BA and clinical relevance for patients with coronary bifurcation lesions treated by drug-eluting stent （DES）. Methods BA was calculated by 3-D quantitative coronary analysis from 347 patients in DKCRUSH-II study. Primary endpoint was the occurrence of composite major adverse cardiac events （MACE） at 12-month, including cardiac death, myocardial infarction （MI） and target vessel revascularization （TVR）. Secondary end points were the rate of binary restenosis and stent thrombosis at 12-month. Results Stenting was associated with the reduction of distal BA. The cut-off value of distal BA for predicting MACE was 60°. Distal BA in 〈60° group had less reduction after stenting （（-1.96＋13.58）°vs. （-12.12±23.58）°, P 〈0.001）; two-stent technique was associated with significant reduction of distal BA （△（-4.05±14.20）°）, compared to single stent group （4＋1.55±11.73, P=0.003）; the target lesion revascularization （TLR）, TVR and MACE rate was higher in one-stent group （16.5%, 19.0% and 21.5%）, compared to two-stent group （3.8%, P=0.002; 7.5%, P=0.016; and 9.8%, P=0.024）, respectively. Among patients in ≥60° group, there were no significant differences in distal BA, stent thrombosis （ST）, MI, MACE, death, TLR, TVR between one- and two-stent groups; after stenting procedure, there was only slight change of distal BA in left anterior descending （LAD）-Ieft circumflex （LCX） subgroup （from （88.54±21.33）° at baseline to （82.44±31.72）° post-stenting）, compared to either LAD-diagonal branch （Di）, or LCX-obtuse marginal branch （OM）, or IRCA distal （RCAd）（all P 〈0.001 ）. Conclusion Two-stent technique was associated with significant reduction of distal BA. DK crush stenting had reduced rate of MACE in patients in 〈60° group, compared to one-stent technique.

关键词： coronary bifurcation lesions bifurcation angle major adverse cardiac events roc curve

Neural network-based computer-aided diagnosis in distinguishing malignant from benign solitary pulmonary nodules by computed tomography

在线全文

学校读者我要写书评

暂无评论

Chinese Medical Journal 2007年第14期120卷 1211-1215页

作者： CHEN Hui WANG Xiao-hua MA Da-qing MA Bin-rong Institute of Biomedical Engineering Capital Medical University Beijing 100069 China Department of Radiology Beijing Friendship Hospital Capital Medical University Beijing 100053 China

Background Computer-aided diagnosis （CAD） of lung cancer is the subject of many current researches. Statistical methods and artificial neural networks have been applied to more quantitatively characterize solitary pulmonary nodules （SPNs）. In this study, we developed a CAD scheme based on an artificial neural network to distinguish malignant from benign SPNs on thin-section computed tomography （CT） images, and investigated how the CAD scheme can help radiologists with different levels of experience make diagnostic decisions. Methods Two hundred thin-section CT images of SPNs with proven diagnoses （135 small peripheral lung cancers and 65 benign nodules） were analyzed. Three clinical features and nine CT signs of each case were studied by radiologists, and the indices of qualitative diagnosis were quantified. One hundred and forty nodules were selected randomly to form training samples, on which the neural network model was built. The remaining 60 nodules, forming test samples, were presented to 9 radiologists with 3-20 years of clinical experience, accompanied by standard reference images. The radiologists were asked to determine whether a nodule was malignant or benign first without and then with CAD output. Diagnostic performance was evaluated by receiver operating characteristic （roc） analysis. Results CAD outputs on test samples had higher agreement with pathological diagnoses （Kappa=0.841, P〈0.001）. Compared with diagnostic results without CAD output, the average area under the roc curve with CAD output was 0.96 （P〈0.001） for junior radiologists, 0.94 （P=0.014） for secondary radiologists and 0.96 （P=0.221） for senior radiologists, respectively. The differences in diagnostic performance with CAD output among the three levels of radiologists were not statistically significant （P=0.584, 0.920 and 0.707, respectively）. Conclusions This CAD scheme based on an artificial neural network could improve diagnostic performance and assist radiologists in distinguishing malignant from benign SPNs on thin-section CT

关键词： diagnosis, computer-assisted neural networks （computer） solitary pulmonary nodules computed tomography roc curve

Extraction of items identifying hiesho(cold disorder)and their utility in young males and females

在线全文

学校读者我要写书评

暂无评论

Journal of Integrative Medicine 2016年第1期14卷 36-43页

作者： Shunji Sakaguchi Hiroshi Kuge Hidetoshi Mori Junji Miyazaki Tim Hideaki Tanaka KazuyoHanyu Taro Takeda Kazuro Sasaki Research Division of Acupuncture and Moxibustion Kansai University of Health Sciences Osaka 590-0482 Japan Department of Anesthesiology Osaka Medical College Hospital Osaka 569-8686 Japan Faculty of Health Sciences Course of Acupuncture and Moxibustion Tsukuba University of TechnologyTsukuba 305-0821 Japan Takarazuka University of Medical and Health Care Takarazuka 666-0162 Japan Graduate School of Life and Environmental Sciences University of Tsukuba Tsukuba 305-8572 Japan Department of Clinical Acupuncture and Moxibustion Meiji University of Integrative Medicine Kyoto629-0301 Japan Department of Acupuncture Medicine Faculty of Oriental Medicine Suzuka University of Medical Science Suzuka 510-0293 Japan

OBJECTIVE： Hie （cold sensation） is one of the most common health complaints in Japan. Those who suffer from severe hie are considered as having hiesho （cold disorder）. However, exact hiesho symptoms have not been defined clearly and the decision as to whether a person suffers from hiesho is subjective and based on self-awareness. The study was conducted in attempt to develop a standardized hiesho diagnostic scale. METHODS： Subjects comprised 1 146 students. From the self-awareness of hiesho symptoms, males and females were divided into hiesho and non-hiesho groups. Physical, behavioral and adaptive characteristics were compared using the 24-item questionnaire （four-grade survey） and indicators for hiesho symptoms were extracted. Based on the scores, a receiver operating characteristic curve was drawn for the total ordinal scale score of the extracted items in relation to the presence and absence of hiesho symptoms, and an optimal cutoff value was determined. RESULTS： The self-awareness of having hiesho was found in 23.2% males and in 55.6% females. The sensitivity was 84.5% for males and 83.3% for females in the hiesho groups, and the specificity was 86.0% for males and 85.2% for females in the non-hiesho groups. CONCLUSION： A questionnaire consisting of the extracted items may be useful to identify hiesho in young males and females with a high level of accuracy.

关键词： cold hypersensitivity questionnaires roc curve

Diagnostic value of liquid-based cytology and smear cytology in pancreatic endoscopic ultrasound-guided fine needle aspiration: A meta-analysis

在线全文

学校读者我要写书评

暂无评论

World Journal of Clinical Cases 2020年第14期8卷 3006-3020页

作者： Hang-Hai Pan Xin-Xin Zhou Fei Zhao Hui-Yan Chen Yu Zhang Department of Gastroenterology Zhejiang Provincial People's HospitalPeople's Hospital of Hangzhou Medical CollegeHangzhou 310014Zhejiang ProvinceChina Department of Gastroenterology The First Affiliated HospitalCollege of MedicineZhejiang UniversityHangzhou 310003Zhejiang ProvinceChina School of Laboratory Medicine and Life Sciences Wenzhou Medical UniversityWenzhou 325027Zhejiang ProvinceChina

BACKGROUND Smear cytology(SC)using endoscopic ultrasound-guided fine needle aspiration(EUS-FNA)is the established and traditional choice for diagnosing pancreatic lesions.Liquid-based cytology(LBC)is a novel alternative cytological method,however,the comparative diagnostic efficacy of LBC remains inconclusive.AIM To examine the diagnostic efficacy of LBC and SC for pancreatic specimens obtained through EUS-FNA via a systematic review and meta-analysis.METHODS A systematic literature search was performed using PubMed,EMBASE,the Cochrane Library,and Web of Science.The numbers of true positives,false positives,true negatives,and false negatives for each cytological test(LBC and CS)were extracted from the included studies.The pooled sensitivity and specificity and the area under the summary receiver operating characteristic curve(AUC)were calculated,and the AUC was compared by Tukey's multiple comparisons test.The quality of the included studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies II tool.RESULTS A total of 1656 patients in eight studies were included.The pooled sensitivity and specificity and the AUC for LBC were 0.76(95%CI:0.72-0.79),1.00(95%CI:0.98-1.00),and 0.9174,respectively,for diagnosing pancreatic lesions.The pooled estimates for SC were as follows:Sensitivity,0.68(95%CI:0.64-0.71);specificity,0.99(95%CI:0.96-100.00);and AUC,0.9714.Similarly,the corresponding values for LBC combined with SC were 0.87(95%CI:0.84-0.90),0.99(95%CI:0.96-1.00),and 0.9894.Tukey’s multiple comparisons test was used to compare the sensitivities and AUCs of the three diagnostic methods;statistically significant differences were found between the three methods,and LBC combined with SC was superior to both LBC(P<0.05)and SC(P<0.05).The pooled sensitivity and AUC did not change significantly in the sensitivity analysis.CONCLUSION LBC may be sensitive than SC in the cytological diagnosis of pancreatic lesions,however,the superior diagnostic performance of their combination emphasizes their integrated

关键词： Liquid-based cytology Smear cytology Pancreatic lesions Endoscopic ultrasound-guided fine needle aspiration Cytological diagnosis roc curve

Combined detection of serum tumor markers for differential diagnosis of solid lesions located at the pancreatic head

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Hepatobiliary & Pancreatic Diseases International 2007年第6期6卷 641-645页

作者： Liao, Quan Zhao, Yu-Pei Yang, Ying-Chi Li, Li-Jun Long, Xiao Han, Shao-Mei Beijing Union Med Coll Hosp Dept Gen Surg Beijing 100730 Peoples R China

BACKGROUND: The differential diagnosis of solid lesions located at the pancreatic head is very important for choosing therapies and setting up surgical tactics. This study was designed to evaluate the clinical significance of combined measurement of multiple serum tumor markers and the application of the receiver-operating characteristic (roc) curves in the differential diagnosis of solid lesions located at the pancreatic head. METHODS: The serum levels of CA19-9, CA242, CA50 and carcinoembryonic antigen (CEA) in 112 patients with carcinoma of the pancreatic head and 38 patients with focal chronic pancreatitis in the pancreatic head were measured with ELISA. The sensitivity, specificity, positive likelihood ratio (PLR) and negative likelihood ratio (NLR) of the four serum tumor markers were calculated. The roc curves for the four serum tumor markers were constructed and the area under the curve (AUC) was calculated. RESULTS: The AUCs of CA19-9, CA242, CA50 and CEA were 0.805, 0.749, 0.738 and 0.705; the PLRs were 1.91, 3.43, 5.09 and 5.46; and the NLRs were 0.41, 0.56, 0.59 and 0.71, respectively. Combined measurements increased the diagnostic specificity, and parallel combined testing increased the diagnostic sensitivity. CONCLUSIONS: Combined measurement of serum tumor markers CA19-9, CA242, CA50 and CEA is valuable in differential diagnosis of solid lesions located at the pancreatic head, and CA19-9 has the best diagnostic ability. Combined measurements can increase the specificity of diagnosis. Evaluation with the roc curve is better than the sensitivity or specificity alone and the results are more integrated and objective.

关键词： pancreatic neoplasm tumor markers roc curve differential diagnosis

BP neural networks and random forest models to detect damage by Dendrolimus punctatus Walker

在线全文

学校读者我要写书评

暂无评论

Journal of Forestry Research 2020年第1期31卷 107-121页

作者： Zhanghua Xu Xuying Huang Lu Lin Qianfeng Wang Jian Liu Kunyong Yu Chongcheng Chen College of Environment and Resources Fuzhou UniversityFuzhou 350116People’s Republic of China Fujian Provincial Key Laboratory of Resources and Environment Monitoring and Sustainable Management and Utilization Sanming 365004People’s Republic of China Key Lab of Spatial Data Mining and Information Sharing Ministry of EducationFuzhou 350116People’s Republic of China

The construction of a pest detection algorithm is an important step to couple"ground-space"characteristics,which is also the basis for rapid and accurate monitoring and detection of pest damage.In four experimental areas in Sanming City,Jiangle County,Sha County and Yanping District in Fujian Province,sample data on pest damage in 182 sets of Dendrolimus punctatus were collected.The data were randomly divided into a training set and testing set,and five duplicate tests and one eliminating-indicator test were done.Based on the characterization analysis of the host for D.punctatus damage,seven characteristic indicators of ground and remote sensing including leaf area index,standard error of leaf area index(SEL)of pine forest,normalized difference vegetation index(NDVI),wetness from tasseled cap transformation(WET),green band(B2),red band(B3),near-infrared band(B4)of remote sensing image are obtained to construct BP neural networks and random forest models of pest levels.The detection results of these two algorithms were comprehensively compared from the aspects of detection precision,kappa coefficient,receiver operating characteristic curve,and a paired t test.The results showed that the seven indicators all were responsive to pest damage,and NDVI was relatively weak;the average pest damage detection precision of six tests by BP neural networks was 77.29%,the kappa coefficient was 0.6869 and after the RF algorithm,the respective values were 79.30%and 0.7151,showing that the latter is more optimized,but there was no significant difference(p>0.05);the detection precision,kappa coefficient and AUC of the RF algorithm was higher than the BP neural networks for three pest levels(no damage,moderate damage and severe damage).The detection precision and AUC of BP neural networks were a little higher for mild damage,but the difference was not significant(p>0.05)except for the kappa coefficient for the no damage level(p<0.05).An"over-fitting"phenomenon tends to occur in BP neural networks,while RF method is more robust,provi

关键词： BP neural networks Detection precision Kappa coefficient Pine moth Random forest roc curve

Selecting decision trees for power system security assessment

在线全文

学校读者我要写书评

暂无评论

能源与人工智能(英文)

能源与人工智能(英文) 2021年第4期6卷 21-30页

作者： Al-Amin B.Bugaje Jochen L.Cremer Mingyang Sun Goran Strbac Department of Electrical&Electronic Engineering Imperial College LondonLondonSW72AZUK Department of Electrical Sustainable Energy TU DelftMekelweg 52628 CD DelftNetherlands Department of Control Science and Engineering Zhejiang UniversityHangzhou 310027China

Power systems transport an increasing amount of electricity,and in the future,involve more distributed renewables and dynamic interactions of the equipment.The system response to disturbances must be secure and predictable to avoid power blackouts.The system response can be simulated in the time domain.However,this dynamic security assessment(DSA)is not computationally tractable in real-time.Particularly promising is to train decision trees(DTs)from machine learning as interpretable classifiers to predict whether the systemwide responses to disturbances are secure.In most research,selecting the best DT model focuses on predictive accuracy.However,it is insufficient to focus solely on predictive accuracy.Missed alarms and false alarms have drastically different costs,and as security assessment is a critical task,interpretability is crucial for operators.In this work,the multiple objectives of interpretability,varying costs,and accuracies are considered for DT model selection.We propose a rigorous workflow to select the best classifier.In addition,we present two graphical approaches for visual inspection to illustrate the selection sensitivity to probability and impacts of disturbances.We propose cost curves to inspect selection combining all three objectives for the first time.Case studies on the IEEE 68 bus system and the French system show that the proposed approach allows for better DT-selections,with an 80%increase in interpretability,5%reduction in expected operating cost,while making almost zero accuracy compromises.The proposed approach scales well with larger systems and can be used for models beyond DTs.Hence,this work provides insights into criteria for model selection in a promising application for methods from artificial intelligence(AI).

关键词： Dynamic security assessment Machine learning Decision trees roc curve Cost curves Cost sensitivity