检索结果-内蒙古大学图书馆

Two-part models for assessing misrepresentation on risk status

EUROPEAN ACTUARIAL JOURNAL 2021年第2期11卷 503-539页

作者： Chen, Li-Chieh Su, Jianxi Xia, Michelle Purdue Univ Dept Stat 150 N Univ St W Lafayette IN 47906 USA Northern Illinois Univ Dept Stat & Actuarial Sci De Kalb IL 60115 USA

Claims modeling is a classical actuarial task aimed to understand the claim distribution given a set of risk factors. Yet some risk factors may be subject to misrepresentation, giving rise to bias in the estimated risk effects. Motivated by the unique characteristics of real health insurance data, we propose a novel class of two-part aggregate loss models that can (a) account for the semi-continuous feature of aggregate loss data, (b) test and adjust for misrepresentation risk in insurance ratemaking, and (c) incorporate an arbitrary number of correctly measured risk factors. The unobserved status of misrepresentation is captured via a latent factor shared by the two regression models on the occurrence and size of aggregate losses. For the complex two-part model, we derive explicit iterative formulas for the expectation maximization algorithm adopted in parameter estimation. Analytical expressions are obtained for the observed Fisher information matrix, ensuring computational efficiency in large-sample inferences on risk effects. We perform extensive simulation studies to demonstrate the convergence and robustness of the estimators under model misspecification. We illustrate the practical usefulness of the models by two empirical applications based on real medical claims data.

关键词： expectation maximization algorithm Loss modeling Misrepresentation fraud Mixture models Semi-continuous data

来源：评论

学校读者我要写书评

暂无评论

Camera motion estimation using particle filters

Camera motion estimation using particle filters

引用

3rd International Conference on Computer Vision Theory and Applications

作者： Nikitidis, Symeon Zafeiriou, Stefanos Pitas, Ioannis Aristotle Univ Thessaloniki Dept Informat Thessaloniki 54124 Greece

ISBN: (纸本)9789898111210

In this paper a novel algorithm for estimating the parametric form of the camera motion is proposed. A novel stochastic vector field model is presented which can handle smooth motion patterns derived from long periods of stable camera movement and also can cope with rapid motion changes and periods where camera remains still. A set of rules for robust and online updating of the model parameters is also proposed, based on the expectation maximization algorithm. Finally, we fit this model in a particle filters framework, in order to predict the future camera motion based on current and prior knowledge.

关键词： camera motion estimation vector field model particle filtering expectation maximization algorithm

来源：评论

学校读者我要写书评

暂无评论

Projection-based Scanned Image Enhancement

Projection-based Scanned Image Enhancement

引用

24th Internationational Conference on Digital Printing Technologies/Digital Fabrication 2008

作者： Ahmed, Mohamed Nooman Lexmark Int Inc Software Res Lexington KY USA

ISBN: (纸本)9780892082797

In this paper, we present a new system to segment and label document images by combining statistical and multiscale view of different image components. Texture of text, halftone and images are characterized by modeling the distribution of a novel intensity projection technique using a mixture of K Gaussians. Model parameters are then estimated using the expectation maximization (EM) algorithm. Using the proposed algorithm, halftone areas were successfully differentiated from text regions

关键词： expectation maximization algorithm

来源：评论

学校读者我要写书评

暂无评论

Noise adaptive speech recognition based on sequential noise parameter estimation

Noise adaptive speech recognition based on sequential noise ...

引用

Workshop on Adaptation Methods for Speech Recognition

作者： Yao, KS Paliwal, KK Nakamura, S ATR Spoken Language Translat Res Labs Kyoto Japan Griffith Univ Sch Microelect Engn Brisbane Qld Australia

In this paper, a noise adaptive speech recognition approach is proposed for recognizing speech which is corrupted by additive non-stationary background noise. The approach sequentially estimates noise parameters, through which a nonlinear parametric function adapts mean vectors of acoustic models. In the estimation process, posterior probability of state sequence given observation sequence and the previously estimated noise parameter sequence is approximated by the normalized joint likelihood of active partial paths and observation sequence given the previously estimated noise parameter sequence. The Viterbi process provides the normalized joint-likelihood. The acoustic models are not required to be trained from clean speech and they can be trained from noisy speech. The approach can be applied to perform continuous speech recognition in presence of non-stationary noise. Experiments conducted on speech contaminated by simulated and real non-stationary noise show that when acoustic models are trained from clean speech, the noise adaptive speech recognition system provides improvements in word accuracy as compared to the normal noise compensation system (which assumes the noise to be stationary) in slowly time-varying noise. When the acoustic models are trained from noisy speech, the noise adaptive speech recognition system is found to be helpful to get improved performance in slowly time-varying noise over a system employing multi-conditional training. (C) 2003 Elsevier B.V. All rights reserved.

关键词： noisy speech recognition non-stationary noise expectation maximization algorithm Kullback proximal algorithm

来源：评论

学校读者我要写书评

暂无评论

Real-time reliability evaluation based on damaged measurement degradation data

引用

Journal of Central South University 2012年第11期19卷 3162-3169页

作者：王小林蒋平郭波程志君 College of Information Systems and Management National University of Defense Technology

A method was proposed to evaluate the real-time reliability for a single product based on damaged measurement degradation *** researches on degradation analysis often assumed that the measurement process did not have any impact on the product's ***,in some cases,the measurement process may exert extra stress on products being *** obtain trustful results in such a situation,a new degradation model was ***,by fusing the prior information of product and its own on-line degradation data,the real-time reliability was evaluated on the basis of Bayesian *** make the proposed method more practical,a procedure based on expectation maximization (EM) algorithm was presented to estimate the unknown ***,the performance of the proposed method was illustrated by a simulation *** results show that ignoring the influence of the damaged measurement process can lead to biased evaluation results,if the damaged measurement process is involved.

关键词： degradation analysis damaged measurement real-time reliability expectation maximization algorithm

来源：评论

学校读者我要写书评

暂无评论

Piecewise polynomial high-order hidden Markov models with applications in speech recognition 16

Piecewise polynomial high-order hidden Markov models with ap...

引用

IEEE International Conference on Computer and Information Technology (CIT)

作者： Lee, Lee-Min Da Yeh Univ Dept Elect Engn Changhua Taiwan

ISBN: (纸本)9781509043149

The assumptions on the hidden Markov model will limit the output of a model to be a piecewise stationary random sequence that may not be a good fit for real processes. In this paper we propose a piecewise polynomial high-order hidden Markov model so that the output of a model can be more versatile. We derived formulas for the calculation of the probability that a given sequence is produced by a model. We also derived the posterior probability of the states and the state transitions and used them in an expectation maximization algorithm to update the model parameters. Experiments on speech recognition of Mandarin digits were conducted to investigate the effectiveness of the proposed model. Experimental results showed that the proposed model can reduce the recognition error rate compared to a baseline hidden Markov model.

关键词： speech recognition expectation maximization algorithm piecewise polynomial high-order hidden Markov model

来源：评论

学校读者我要写书评

暂无评论

Considerations on regulatory sequences of the distal promoter region of the HLA-G gene

引用

HUMAN IMMUNOLOGY 2013年第4期74卷 473-477页

作者： da Silva, Jose Samuel Slowik, Renata Bicalho, Maria da Graca Univ Fed Parana UFPR Lab Imunogenet & Histocompatibilidade LIGH Dept Genet Ctr Politecn BR-81530990 Curitiba Parana Brazil

Gene expression in eukaryotic cells is accomplished via association of transcription factors, some of which directly bind to DNA regulatory sequences. HLA-G codes for an immunoregulatory protein with tissue-specific expression, its unique promoter regulatory region is responsible for this feature. The aim of the present study was to explore motif composition as well as identify haplotypes in the HLA-G 5' distal promoter region. The sample was composed by 176 euro-descendents individuals genotyped by Sequence Based Typing of HLA-G distal promoter, encompassing 16 SNPs. Haplotypes were inferred by the expectation maximization algorithm. Only haplotypes with frequency higher than 1% were aligned to check for similarities and differences and thirteen haplotypes remained. For a better understanding of the nucleotide diversity of the analyzed region our approach was to split the whole sequence into two regions. Two contrasting haplotype groups were found in both regions, allowing us to suggest the existence of different transcription factors capable of binding cis elements while the intra-group variations suggest the intensity modulation of binding with regulatory factors. (C) 2012 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.

关键词： CIITA class II trans activator EM expectation maximization algorithm GAS interferon-gama activated site HLA human leucocyte antigen HSE heat shock element HSF-1 heat shock factor 1 HWF>1% haplotypes with frequency higher than 1% INF interferon ISRE interferon stimulated regulatory element LCR locus control region MHC major histocompatibility complex SNP single nucleotide polymorphism URR upstream regulated region UTR untranslated region

来源：评论

学校读者我要写书评

暂无评论

Cheating Detection Method Based on Improved Cognitive Diagnosis Model 18th

Cheating Detection Method Based on Improved Cognitive Diagno...

引用

18th Annual International Conference on Web-Based Learning (ICWL)

作者： Li, Zhizhuang Zhu, Zhengzhou Xie, Qiongyu Peking Univ Sch Software & Microelect Beijing Peoples R China Anhui Normal Univ Wuhu Anhui Peoples R China

ISBN: (纸本)9783030357580;9783030357573

Cheating in examinations destroys the principles of fairness and justice in evaluation. Cheating detection is of great practical significance. Traditional cheating detection methods have many disadvantages, such as difficult to detect covert equipment cheating, multi-source cheating, difficult to distinguish plagiarists from plagiarists, difficult to distinguish plagiarists from victims, or plagiarism from coincidences. In this paper, the concept of knowledge point mastery Index is introduced to measure students' mastery of a certain knowledge point, and a test method of cheating based on improved cognitive diagnostic model is proposed. This method calculates the weight of each knowledge point in every examination question through linear regression and EM algorithm according to students' historical learning behavior, and then calculates students' mastering degree of knowledge point based on historical answers. Then calculate the mastering degree of knowledge point based on the examination results. Finally, we compare the mastering degree of knowledge point based on the examination results and the historical answers to detect students' cheating situation. The experiments show that the precision and recall rate of this method are significantly higher than those of the method based on the false-same rate, the method based on the false-same rate and the right-same rate and the method based on the Person-Fit index.

关键词： Weight of knowledge point Cheating detection Mastering degree of knowledge point Linear regression expectation maximization algorithm

来源：评论

学校读者我要写书评

暂无评论

Identification of ARX System Based on Shifted Asymmetric Laplace Distribution 38

Identification of ARX System Based on Shifted Asymmetric Lap...

引用

38th Chinese Control Conference (CCC)

作者： Yu, Miao Zhang, Tianyi Yang, Xianqiang Harbin Inst Technol Res Inst Intelligent Control & Syst Harbin Peoples R China

ISBN: (纸本)9789881563972

The identification of AutoRegressive eXogenous (ARX) model by outliers is addressed in this paper. Shifted(non-centralized) asymmetric Laplace (SAL) distribution and expectation maximization (EM) algorithm are employed to estimate the unknown model parameters. Outliers are common in the signal acquisition process and have a serious impact on data-driven modeling method. In this paper, the probability method is used to solve the problem of outliers. When the noise parameter is regarded as a prior exponential distribution, the model output obeys the SAL distribution which is robust to outliers. The known statistical properties of SAL distribution are applied to calculate the M-step in the EM algorithm and get the iterative parametric formula. The accuracy of the proposed algorithm is verified by a numerical simulation example.

关键词： shifted asymmetric Laplace distribution expectation maximization algorithm autoRegressive eXogenous model

来源：评论

学校读者我要写书评

暂无评论

MaxMin-RLHF: Alignment with Diverse Human Preferences 41

MaxMin-RLHF: Alignment with Diverse Human Preferences

引用

41st International Conference on Machine Learning (ICML)

作者： Chakraborty, Souradip Qiu, Jiahao Yuan, Hui Koppel, Alec Manocha, Dinesh Huang, Furong Bedi, Amrit Singh Wang, Mengdi Univ Maryland Dept Comp Sci College Pk MD 20742 USA Princeton Univ Dept Elect & Comp Engn Princeton NJ 08544 USA JP Morgan Chase AI Res New York NY USA Univ Cent Florida Dept Comp Sci Orlando FL 32816 USA

Reinforcement Learning from Human Feedback (RLHF) aligns language models to human preferences by employing a singular reward model derived from preference data. However, the single reward model overlooks the rich diversity of human preferences inherent in data collected from multiple users. In this work, we first derive an impossibility result of alignment with single reward RLHF, thereby highlighting its insufficiency in representing diverse human preferences. Next, we propose to learn a mixture of reward models via an expectation-maximization algorithm and solve a MaxMin alignment objective inspired by the Egalitarian principle in social choice theory to better honor diverse human preferences. We present comprehensive experimental results on small-scale (GPT-2) and large-scale language (with Tulu2-7B)) and show the efficacy of the proposed approach in the presence of diversity among human preferences. We remark that our findings in this work are not only limited to language models but also extend to reinforcement learning in general.

关键词： expectation maximization algorithm

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：