最新刊期
卷 45 ， 期 4 ， 2017
- 摘要：The traditional microblog community detection algorithm has the characteristic of low coupled clustering and the overlapping degree can not be controlled.In this paper,we present a divisive approach for overlapping microblog community detection algorithm via core tags.Firstly,the key idea is to develop a tag weighing strategy by taking advantage of the co-occurrence of tags and inverse user frequency.Then tag correlation can be exploited,which investigates both inter and intra correlation of tags,and the tags for users can therefore be expanded.Users containing certain tag in the whole community are extracted as a temporary group and the quality value is calculated under the current partition.The most appropriate core tag is selected and the corresponding group is then updated until certain requirements are satisfied.The community detected by this algorithm share common core tags and the partition results can be revised based on the explicit and implicit interest of users,together with the users' attention and practical application.Experimental results show that the method is effective and has practical significance.关键词：microblog network;overlapping community detection;core tag;user attention relationship;tag cut190|940|4更新时间：2025-07-16
- 摘要：In order to adapt to the requirements of the business process multi-functional and dynamic change,getting business processes with modular substitution is one of the effective ways to solve these problems.The concept of differential control Petri net and differential data Petri net are proposed based on open Petri net and differential Petri net.Liveness detection in control flow and data flow of models is examined by the evolution expression of differential Petri net through semantics,and stability analysis of substitution module and its discrete and continuous parts are done by differential Petri net,respectively,to achieve business process module adaptation analysis.Theoretical analysis results show that the established model of differential expression can reflect the liveness and stability of models,module adaptation effect is better.Finally,we develop the simulation analysis based on data platform;the experimental results show that the presented method is effective.关键词：differential Petri net;control flow;data flow;module adaptation161|345|1更新时间：2025-07-16
- 摘要：With the development of computer graphics and computer vision technology,light field comes into sight and is rapidly applied in various fields.However,the acquisition of the light field needs a large amount of pictures,which has the characteristics of large data and high cost.So how to use a small amount of data to obtain the light field has been paid more and more attention,and a lot of work has been done.To address the above problems,a new method of light field reconstructing is proposed,which combines with wavelet transform and sparse Fourier transform by using sparseness of light field in angle domain.First,we use multi-resolution analysis characteristic of wavelet transform,and the original image can be decomposed into four sub-frequency images through wavelet transform.Then the frequency positions of four sub-frequencies are separately recovered through the Fourier slice theorem,and their two-dimensional angle spectrum are further obtained,respectively.Finally,the light field is obtained by combining the two-dimensional angle spectrum of each sub-frequency image and making inverse wavelet transform.In the proposed method,the original image is decomposed into four sub-images by using wavelet transform,and the sub-images are reconstructed respectively.This not only reduces the complexity of our method and greatly reduces the running time of our method,which provides the basis for the wide application of the light field,but also our method effectively inhibits the window effect by comparing to only using sparse Fourier algorithm,so that the reconstruction result is more accurate.In addition,the method can effectively improve problem of small frequency leakage in off-grid recovery by separating high frequency and low frequency information,and further improves the reconstruction results.In the end,the effectiveness of the algorithm is verified by simulation.关键词：wavelet transform;light field reconstruction;sparse;Fourier transform;window effect376|517|3更新时间：2025-07-16
- 摘要：A golden section Fibonacci tree optimization algorithm for multimodal function optimization is proposed,which is based on the Fibonacci tree optimization algorithm (FTO) and is combined with the golden section theory.By making full use of the features upon the global and local alternation optimization of Fibonacci tree optimization,the algorithm has improved and enhanced its local searching ability and small peak searching ability through the method of golden section ratio compression to optimize problems of solution space.Simulation result of typical test function shows that the algorithm owns strong optimization capacity,fast speed as well as high precision.关键词：multimodal function optimization;Fibonacci tree optimization;golden section;global and local alternation430|658|4更新时间：2025-07-16
- 摘要：In order to improve the efficiency of tracking algorithm based on distribution fields and the robustness of the algorithm under complex background,tracking algorithm by compressive distribution fields with adaptive hierarchical structure is presented.Distribution of pixel values in target region is considered in this method,k-means algorithm is introduced to analyse the distribution of pixel values in the first frame,adaptive hierarchical structure of distribution fields is built according to the clustering results.For the problem that the dimension of distribution field model is high,compressive sensing is combined to compress distribution fields,which can reduce the model dimension and improve the efficiency of tracking algorithm.Furthermore,local search strategy in original distribution fields tracking algorithm is changed,random sampling is used to improve the tracking accuracy.Experimental results demonstrate that the proposed algorithm outperforms the state-of-the-art tracking algorithms.关键词：distribution fields;compressive sensing;object tracking;cluster analysis122|360|1更新时间：2025-07-16
- 摘要：Since it is difficult to obtain adequate performance distribution information of highly reliable components in case of related competing failures and system reliability problems cannot be accurately estimated,the multi-state system reliability evaluation method in consideration of epistemic uncertainty was proposed in case of related competing failures.Firstly,the components' sudden failure thresholds were assumed as a decreasing random process to represent the correlation between accumulative degradation and sudden failure.Secondly,in order to reduce the impact on the epistemic uncertainty of components,the components' performance distribution model based on interval variables was built on the assumption that the components' performance damage distribution parameter and sudden failure parameter caused by random shocks are interval variables.Thirdly,the traditional general generating function method was improved and the definition of general generating function and its algorithm were proposed.Finally,the reliability of the power amplifier of a radar system was analyzed.With simple and clear procedures,this method not only avoids the components' complex failure mode and less status information,but also features great universality and high engineering application value.关键词：related competing failure;epistemic uncertainty;interval-valued universal generating function method;power amplifier system;reliability91|733|4更新时间：2025-07-16
- 摘要：In order to reduce the negative influence of outliers on the model of support vector data description (SVDD) when the training dataset contains both normal samples and outliers which are all labeled as target class,a one-cluster kernel possibilistic C-means based SVDD method for outlier detection is proposed.In this paper,each sample of the training dataset is assigned a confidence level based on the membership degree of each sample belonging to the normal class,which is obtained through the one-cluster kernel PCM clustering.The proposed algorithm incorporates the confidence levels into the training model to reduce the importance of the samples which have less confidence levels.The experimental results show that the proposal significantly improves the effect of outlier detection,compared with the existing SVDD-based outlier detection methods.关键词：outlier detection;support vector data description;possibilistic C-means;confidence level166|387|6更新时间：2025-07-16
- 摘要：Two direct TDOA geolocation approaches are proposed for FH(Frequency-Hopping) signals-the maximum likelihood direct geolocation and the maximum correlation-sum direct geolocation.The first approach makes use of the sparsity of FH signals in frequency domain,and constructs CCF(cross correlation function) matrix in frequency domain,the location estimate is obtained by searching the max eigenvalue of CCF in a two dimensional grid.The second approach directly uses the sum of CCF,and searches in a two dimensional gird to find the location estimate.The comparison between them has been made,and the first approach performs better than the second one,but needs much more computation.Monte Carlo simulations validate that the accuracy of the proposed maximum likelihood direct geolocation method is the best,and the maximum correlation-sum direct geolocation method also outperforms the two-step method.关键词：frequency-hopping;TDOA;direct geolocation168|344|4更新时间：2025-07-16
- 摘要：It's a key problem to find wavelet invariants to the transformation of scale,translation and rotation in pattern recognition using multi-resolution analysis.Moment invariant is a mature method on theory and applications.This paper links the moment invariants with the approximation coefficients of image wavelet decomposition.A novel biorthogonal wavelet moment invariant is derived from the biorthogonality of the spatial basis functions.The experimental results are also provided to confirm the correctness of the theoretical derivation.After that,the limiting condition of the conclusion is analyzed by taking Haar wavelet as an example.Both theoretical analysis and experimental verification show that the wavelet moments of higher order than smoothness can be calculated within required accuracy.And the complete theoretical and experimental results are obtained.Finally,some problems to be paid attention to in practical application are pointed out.关键词：pattern recognition;multi-scale analysis;biorthogonal wavelets;invariant moment;smoothness159|829|0更新时间：2025-07-16
- 摘要：Passive inter-modulation fault location technology based on the reference signal is researched.The technology has one more reference channel than the traditional passive inter-modulation measurement techniques,to provide a reference signal,to demodulate phase response of the inter-modulation signal,while introducing calibration technology in the measurement port to build the reference plane,Obtained Time domain response relative to the reference plane By the time-frequency conversion and effective troubleshooting.Based on this,the fault location test is simulated to prove the validity of theory,and the phase reference board is designed and fabricated,the experimental results and the actual value showed that:Absolute value of the difference between the actual value and the measured value of two fault location on the cables are ≤ 1m,the actual relative error of approximately 5.1%,2.4%,proves the correctness and feasibility of this approach.The technology will overcome the shortcomings of traditional fault location,adapts to the nasty geographical environment.关键词：passive inter-modulation;fault location;reference signal;calibration technology;reference plane173|381|1更新时间：2025-07-16
- 摘要：The paper proposes a local linear-periodogram-Kalman filtering (LL-P-KF) hybrid algorithm.Firstly,a local linear method is established;secondly,the periods of sinusoidal signal are detected from modeling error by using periodogram;finally,the sinusoidal signal is recovered from the error by using Kalman filtering.Simulation results show that the LL-P-KF hybrid algorithm has better recovery results.关键词：chaotic signal;local linear prediction;weak sinusoidal signal;periodogram;Kalman filtering145|354|3更新时间：2025-07-16
- 摘要：In order to maximize the energy efficiency (EE) of a cognitive small cell base station,this paper analyzes downlink joint spectrum resource blocks (SRBs) and power allocation based on game theory.In an interference-limited environment,base stations can share unused spectrum resources in a distributed architecture.The proposed algorithm introduces power and interference temperature constraints to avoid harmful accumulated interference on primary users.It is non-convex optimal to optimize the fractional form EE under multiple coupling constraints.The maximization problem is transmitted into an equivalent problem in subtractive form which can be solved from the iterative point of view.After obtaining the SRBs allocation strategies,the original game can be re-modeled as equivalent sub-games,then the transmission power can be solved more easily after decoupling power constraints based on pricing.Simulation results show that the proposed algorithm can converge to a Nash equilibrium and effectively improve system resources utilization and EE.关键词：cognitive small cell networks;energy efficiency;interference temperature;spectrum allocation;power allocation146|823|0更新时间：2025-07-16
- 摘要：Aiming at the intrinsic problems of Chinese Remainder Theorem in fountain decoding process with modular arithmetic,this paper proposes a decoding algorithm based on extended Euclidean theorem.The linear congruence equations are merged in the extended Euclidean decoding algorithm,which avoids the failure of solving the rate factor when the decomposition factors are non-coprime.In the modular arithmetic fountain encoding process,the original packet is continuously decomposed by the factor,which is randomly selected from the natural number,into the encoded packets consisting of the residues and the factors.When a certain amount of packets are received,it can be achieved to decode successfully.The codec efficiency has been improved as the algorithm has extended the range of the modular arithmetic factor.Through theoretical analysis and numerical simulation,the effectiveness of this decoding algorithm of modular arithmetic fountain code has been proved.关键词：fountain codes;Chinese remainder theorem;extended Euclidean;remainder transform codes;linear congruence equations257|910|1更新时间：2025-07-16
- 摘要：The integer ambiguity resolution is the key to precise attitude determination.In case of single frequency single epoch,this paper presents a fast integer ambiguity resolution method based on nonlinear constrained integer least-squares solved with the quadratic eigenvalue problem,and applies it to BDS attitude determination system.The method will integrate the baseline length called the priori information into objective function,and construct new objective function to improve the accuracy of integer ambiguity solution.The effectiveness of the proposed method is verified by experiments on the BDS attitude determination platform.The results show that,in the single frequency single epoch BDS attitude determination,the method has highly improved the success rate of integer ambiguity and attitude determination resolution.关键词：BDS;attitude determination;integer ambiguity;nonlinear constrained116|311|2更新时间：2025-07-16
- 摘要：Recently,ensemble Kalman filter is considered as an effective solution for the state estimation of nonlinear system.Aiming at the consistency deviation occurred in virtual measurement sampling process on account of measurement noise uncertainty,a novel multi-sensor ensemble Kalman filtering algorithm based on Metropolis-Hastings sampling is proposed.Firstly,combined with the physical properties of multi-sensor measurement system and the generation mechanism of bootstrapping measurement in ensemble Kalman filter,multi-sensor bootstrapping measurement set is structured.Secondly,through solving the likelihood of multi-sensor bootstrapping measurement and designing the probability function of measurement acceptance,validation measurement from multi-sensor bootstrapping measurement set is confirmed by Metropolis-Hastings sampling strategy.The new method corrects the consistency deviation appearing at bootstrapping measurement by means of the extraction and utilization for the redundancy and complementary information in multi-sensor measurement,and improves the filtering precision for the estimated system state.Finally,the theoretical analysis and experimental results show the feasibility and efficiency of our proposed algorithm.关键词：nonlinear filter;ensemble Kalman filter;bootstrapping measurement;Metropolis-Hastings sampling168|315|0更新时间：2025-07-16
- 摘要：Compressed sensing (CS) has been successfully applied to the synthetic aperture radar (SAR) imaging.These CS-based SAR imaging algorithms generally assume that the model of the imaging system is accurate.However,in practice it is common to encounter model errors which usually introduce unknown phase errors into the acquired data.The phase errors may cause range migration or defocusing.In this paper,an approach for matrix form joint CS-SAR imaging and autofocus is proposed.Based on smoothed l0 norm (SL0) algorithm,we develop a matrix form regularized SL0 (MRSL0) algorithm to efficiently perform CS-SAR imaging.The MRSL0 adopts inequality constrain to tolerate phase errors and has fast computation speed due to its matrix form.Experiment results demonstrate that the proposed approach can efficiently reconstruct high quality images using limited amount of measurements.关键词：synthetic aperture radar (SAR);compressed sensing (CS);smoothed l 0-norm (SL0) algorithm;matrix form regularized SL0 (MRSL0) algorithm178|387|2更新时间：2025-07-16
- 摘要：The coherent accumulation in a single frame usually suffers serious information loss in some cases and it will be even worse through multi-frame incoherent accumulated,influencing the performance of DP-TBD algorithm.A new DP-TBD algorithm is proposed based on information geometry theory which uses the covariance matrix in every resolution cell to implement multi-frame accumulation directly,avoiding the information loss of coherent accumulation in a single frame with better detecting and tracking performance.Finally,the effectiveness of the proposed algorithm is verified by simulation test.关键词：dynamic programming;track before detection;information geometry;statistic manifold;riemannian mean266|376|4更新时间：2025-07-16
- 摘要：To address the problem of changes of business processes for an enterprise or organization,we utilize the normal and exceptional instances to recommend the next possible activity for the current incomplete workflow instance.Since every workflow instance is a sequence of activity names,it cannot be calculated numerically.we firstly extract the order of each activity in the sequence as a number value,and then get a matrix which is similar to User-Item matrix in traditional recommendation systems.This matrix can facilitate the calculation of similarity between two workflow instances.Finally,we choose these complete instances which are most similar to the current incomplete instance,construct the activity list as the recommendation result by these instances.Experimental results show that the proposed algorithm is effective and efficient.关键词：collaborative filtering;activity recommendation;recommendation system;workflow activity;workflow instance189|501|0更新时间：2025-07-16
- 摘要：To recommend useful microblogs that match users' interests and likes effectively,an approach in which the dynamic interests and social networking (DISN) of users are seamlessly integrated based on LDA model is proposed.The approach infers the interest vector of users better by using time function and groups the new published microblogs by clustering method and gets the best matching groups with users' interest vector.Then DISN traverses the selected groups by grid querying approach and matches the microblogs with publishers' probabilities of being followed and sorts the result.Finally the personalized microblogging recommendation is achieved.Experimental results show that DISN is more effective and efficient than the traditional models.关键词：dynamic interests;social networking;LDA;grid querying;personalized recommendation;microblog352|699|12更新时间：2025-07-16
- 摘要：This paper presents an action recognition method based on 3D skeleton.This method redefines the coordinates of the articulations which belong to the skeleton to form a simplistic skeleton model firstly.Then a reformative dynamic time warping (R-DTW) algorithm is applied to implement action recognition.There are no two persons identical in an action owing to the difference of body size,shape and action expression.The simplistic skeleton model could decrease this intra-class variability effectively.The drawbacks of conventional DTW algorithm lie in high computational complexity and low recognition efficiency.To solve this problem,we design a method named "Planning & Refining".We conduct this algorithm on MSR Action3D dataset and the results demonstrate its effectiveness.关键词：skeleton;reformative dynamic time warping(R-DTW);action recognition82|795|4更新时间：2025-07-16
- 摘要：An approach based on composite gradient vector is proposed to overcome the problem of nonlinear distortion in contact fingerprint identification.Firstly,rectangular coordinate system and the first base vector are set up,and then the fingerprint image is rotated.Base vectors are collected on rotated fingerprint image.All the base vectors are set into vector cluster.Then the basis vectors are made into greatly gradient vector.Finally,hierarchical markup method is used for markup on the composite gradient vector.The target fingerprint image is recognized by fingerprint searching,compound gradient vector matching,dimension and gradient matching auxiliary.Experiment results show that the proposed approach has overcome the problem of rotation,translation and scaling situation and has a strong ability of anti-nonlinear deformation with fast recognition speed and high recognition accuracy.关键词：fingerprint matching;nonlinear distortion;composite gradient vector;hierarchical markup130|526|1更新时间：2025-07-16
- 摘要：For the electric vehicle's defects of short continued driving mileage,easy battery aging,poor climbing ability and low energy recovery efficiency,the high power density and long cycling life hybrid power system of the super capacitor-battery based on the super capacitor auxiliary is studied.Based on the structure of the super capacitor-battery system,the driving mode of the electric vehicle is analyzed and the super capacitor charging model is built;meantime,the super capacitor battery hybrid power system is built and simulated.Simulation and experimental results show that the hybrid power system can effectively recover the braking energy,make up the defects of the battery low energy density and low power density,improve the electric vehicle power performance and power utilization.So this system can make the electric vehicle continued driving mileage longer,effectively reduce the battery voltage and current amplitude fluctuations as well as prolong the battery service life.关键词：electric vehicle;super capacitor;battery;the hybrid power system184|971|1更新时间：2025-07-16
- 摘要：Load of microgrid has characteristics of strong randomicity and large fluctuation,so that single point prediction cannot satisfy the need of microgrid operating stability.In this paper,a modified multi-objective optimization prediction intervals(PIs) method for microgrid load is proposed,recurrent neural network (RNN) is adopted to build load prediction model,technique for order preference by similarity to an ideal solution and the grid selection strategy are introduced to modify multi-objective artificial bee colony algorithm(MMOABC),which optimizes the RNN prediction model,improving the accuracy and reliability of microgrid short-term load intervals prediction.The experiment results show that the proposed method for microgrid load has superior performance,which can provide the decision-making basis for the safety and economy of microgrid operation.关键词：microgrid;prediction intervals;recurrent neural network;artificial bee colony70|416|4更新时间：2025-07-16
- 摘要：To solve the existing problem of the cipher processor,high-efficiency reconfigurable block cipher processor based on stream architecture was proposed.Through the efficient data organization and flexible cipher computing units,the processor that adopts the design conception of hierarchy achieves the cooperation of software and hardware pipeline,develops instruction level parallelism in a block and among multiple blocks and improves the utilization rate of functional units.The processor was simulated and synthesized in 65nm CMOS process.The mapping results of typical block cipher algorithms show that it has high encryption performance both in CBC and ECB mode.Compared with other cryptographic processors,this processor has the advantage of small-area and high-efficiency.关键词：block cipher;stream processor;reconfigurable;software pipeline;area efficiency119|339|1更新时间：2025-07-16
- 摘要：A CMOS temperature sensor with digital output is presented,and a new self-correcting technique is proposed to solve the problem of poor consistency and low yield caused by process variations.By applying the self-correcting technique,the change of the reference voltage in the temperature sensor kernel module is effectively suppressed and the consistency among chips is obviously improved.In this paper,the reference voltage module is simulated with different process corners.By comparing the maximum deviation of the reference voltages in the open and close states of the self-correcting module,the effectiveness of self-correcting technique is validated.The complete design is fabricated in the CSMCB5212 0.5 μ m CMOS process,providing an SPI digital interface with 10-bit output data.In the temperature range from -35℃ to 105℃,the inaccuracy of experiment results is merely ±1℃,and the overall power consumption is less than 0.6mW.关键词：temperature sensor;self-correcting;consistency191|961|1更新时间：2025-07-16
- 摘要：Aiming at the problem that the design process for the single-phase full-bridge low frequency inverter is lack of specification,this paper presents a formulation process decoupling method for designing the main circuit filter parameters and the controller parameters in the condition of desired dynamic transients.Based on the resistive load step,the optimal load transient response processes for the single-phase full-bridge low frequency inverter were analyzed in two extreme cases,this paper derived the mathematical formula of the optimal load transient characteristics about filter parameters,and gives an applicable scope of the formula-output voltage undershoot below 10%,and then based on the desired dynamic transients of inverter,a design method of filter parameter is proposed,and take the example of sliding mode control,the design method of sliding mode controller is given.The simulation experiment results show that the optimal load transient response process is existent,and verify the correctness and effectiveness of optimal load transient characteristic expression and the decoupling design method.This decoupling design method has good theoretical and engineering value.关键词：optimal load dynamic transient;filtering;low frequency inverter;single phase;sliding mode control149|516|2更新时间：2025-07-16
- 摘要：Because of the popularity and the change of communication mode for mobile terminals and in order to meet the security access demand of pervasive computing environment,we present a security access model for pervasive computing environment which contains three kinds of access mechanisms:local service domain registration,local domain access and cross-domain access,and the specific work processes of the above access mechanisms are introduced at the same time.The security proof shows that our proposal is provably secure in the CK security model and has superiority in anonymity,security and efficiency,which make it more suitable for pervasive computing environment.关键词：pervasive computing environment;security access model;cross-domain;CK security model69|335|2更新时间：2025-07-16
- 摘要：LBlock-s is the kernel block cipher of the authentication encryption algorithm LAC submitted to CAESAR competition.The general structure of LBlock-s is almost the same as that of LBlock,but LBlock-s adopts an improved key schedule algorithm with better diffusion property.Using the shifting relation of subkeys derived by the key schedule algorithm,an impossible differential cryptanalysis on 21-round LBlock-s was presented based on a 14-round impossible differential.The time and data complexities are 267.61 21-round encryptions and 263 chosen plaintexts respectively,and the number of subkey bits needed to be guessed is 72.Using partial-matching method,an impossible differential cryptanalysis on LBlock-s up to 23-round was also presented with time complexity less than exhaustion of all key bits.This work is useful for the security analysis of LAC algorithm.关键词：LBlock;LBlock-s;key schedule algorithm;impossible differential cryptanalysis305|488|1更新时间：2025-07-16
- 摘要：In order to improve flow table lookup speed,flow characteristics of the OC-192 backbone links was explored.It proves that the backbone traffic not only has high concurrency and high arrival rate,but also has good network locality characteristics in an appropriate cached window.Based on these characteristics and the principle of locality,a fast flow table lookup method was implemented by using a naive Hash table structure with constant increase of auxiliary space.The theoretical analysis and experiments on real-life data traces show that the proposed method can reduce the length and the time of flow table lookup by 20.2% and 17.1% compared with the existing method,respectively.关键词：Hash table;high concurrency network;flow management;network locality185|354|1更新时间：2025-07-16
- 摘要：Data driven software defects prediction is considered as an effective means for the optimization of quality assurance activities.This paper tries to survey the studies of data driven software defects prediction from 2005 to 2015.First,it briefly introduces the research background in this field.Then,it summarizes and analyzes the software modules attribute metrics in detail,and thoroughly surveys the state-of-the-art methods of data driven software defects prediction.The discussed topics include within-release,cross-release and cross-project defects prediction.Finally,the potential research directions in the field are discussed.关键词：software defects prediction;data driven;software metrics;machine learning113|471|8更新时间：2025-07-16
- 摘要：In the field of new type of nonvolatile memory,the simple-structured RRAM with high speed and low consumption has manifested great advantages and competitiveness.In this review,a brief introduction to the structures of RRAM and two kinds of resistive switching behaviors are made.And then,a summary of two kinds of resistive mechanisms is also given.With discussing the performance optimization of RRAM,how the optimized methods achieve balance and unification between the resistive performance and the reliability and stability of the devices is simply demonstrated.In the end,the future of RRAM is also prospected.关键词：RRAM;resistive switching mechanism;performance optimization188|683|2更新时间：2025-07-16
- 摘要：Melody extraction is to automatically extract the melody from a given piece of polyphonic music,and has been widely applied in music retrieval,score transcription,cover song identification and so on.Melody extraction from digital polyphonic music is reviewed in this paper.Firstly,the methods used for melody extraction are classified.Then,the representative methods are elaborated in detail,evaluation methodologies are presented,and the most recently MIREX melody extraction competition results are shown.Finally,the existing challenges are summarized,and the future research directions are provided.关键词：melody extraction;polyphonic music;multi-part music;music signal processing;music information retrieval93|685|2更新时间：2025-07-16
- 摘要：An analytical high-resolution direction-of-arrival (DOA) algorithm based on the estimation of rotation matrix(ERM) is proposed.In order to sufficiently utilize the concerning spatial temporal domain information to improve the performance of the algorithm,a set of correlation matrices of the received signals containing the rotation matrix and possessing diagonal structure is generated.These common complex-valued target matrices are transformed into real symmetrical ones through a series of matrix transformation operations to make use of ACDC to diagonalize the target matrices and to estimate the rotation matrix.The arrival angles could be further derived and the DOA goal could be finally reached.Simulation results illustrate that,compared with other existing state-of -the-art DOA algorithms,the proposed ERM possesses higher resolution and more accurate DOA estimation performance.关键词：direction-of-arrival (DOA) estimation;estimation of rotation matrix (ERM);joint diagonalization (JD);alternating columns diagonal centers (ACDC)138|421|0更新时间：2025-07-16
- 摘要：Sentiment analysis is a very important technology in text mining.However,a number of systems require amounts of annotated training data in different fields.In order to solve these problems,an approach to polarity classification based on sentiment tags is proposed.Firstly,on the basis of all the documents,the sentiment-topic model is developed and the sentiment tags for each review are extracted.Then each review is divided into two sub-texts by these sentiment tags,and each sub-text is classified by exploiting the co-training algorithm.Finally,the category results of two sub-texts are combined to determine document-level polarity of each review.Experimental results show that compared with other algorithms,the method improves the classification precision without a large number of annotated samples.关键词：polarity classification;sentiment tag;semi-supervised learning;co-training learning235|1483|2更新时间：2025-07-16
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0