72 results about "Multi label learning" patented technology

The invention discloses a transfer learning design method and system based on domain adaptation under a multi-example multi-label framework. According to the invention, multi-example multi-label learning and transfer learning are unified into one framework, source domain data samples and target domain data samples are effectively utilized for correlation statistics, and the source domain samples can be effectively used in the learning of target domain tasks; the characteristics of a source domain data sample set and a target domain data set in RKHS are utilized, a two-step domain adaptation process formed by sample weighting and a sample selection mechanism based on clustering is utilized, so that the learning of target tasks has enough training samples weighted and selected from the source domain set; and a miFV algorithm is utilized to convert multi-examples into a single example, and the calculation cost problem of domain adaptation is solved.

The invention discloses a multi-label learning design method based on a hashing method. Through the combination of a hashing algorithm and a multi-label learning algorithm based on Bayesian statistics, the correlation between labels is effectively utilized so as to improve the predicting performance of a multi-label learning model, labels and neighbors of the labels are introduced to computation of the posterior probability through the characteristics of the neighbors, the correlation between the labels is fully considered, and the accuracy of the algorithms is improved; the problem that the label space in multi-label learning of large-scale data is higher in dimension and sparse is solved through an MinHash algorithm; the purpose of learning large-scale data is achieved by finding the neighbors through locality sensitive hashing (LSH), the neighbors can be rapidly and efficiently found, and the expandability of the multi-label learning algorithm is improved.

The invention provides a traditional Chinese medicine diagnosis method and system based on specific weight characteristics of k-nearest neighbor labels to solve the problems. The method includes the steps of acquiring characteristic data weight information of cases of different types according to a preset weight determination strategy; according to the characteristic data weight information, acquiring a weighted Euclidean distance between any two cases, and selecting the preset number of cases having the minimum weighted Euclidean distances; subjecting the selected cases to ML-LSWAKNN (multi-label learning of specific weighted adjustment k-nearest neighbor) to acquire evaluation indexes corresponding to the selected cases. The method and the system have the advantages that the influence of characteristic weighting upon classifying is fully considered, and classifying precision is greatly improved.

The invention provides a Spark big data platform-based CLR multi-label classification method, and relates to a data mining technology. Each data set is divided into an area through a relationship between characteristics and labels; one part of data sets are randomly extracted from the data sets in a training set as test sets; base classifiers are built by the remaining data sets as the training set; and the test sets are used for testing a classifier and modifying parameters of the base classifiers thereof, thereby selecting the best base classifier from multiple trained base classifiers as the base classifier of the data set. A prediction set is predicted by using the final base classifiers. A CLR multi-label learning algorithm is combined with a memory-based efficient computing theory of Spark, no correlation after label transformation in a CLR algorithm is fully utilized, the interference among different base classifiers is reduced, the operating speed of a computing framework of the Spark is fully utilized and data can be effectively mined.

The invention discloses a face attribute recognition method based on a multi-task multi-label learning convolutional neural network, and relates to a computer vision technology. Firstly, multi-task learning is adopted, and two tasks of face key point detection and face attribute recognition are learned at the same time; different learning difficulties and learning convergence rates of different attributes are considered, the attributes are divided into subjective attributes and objective attributes, and the convergence rate of the network is accelerated and the problem of sample imbalance is relieved by adopting a dynamic weight and an adaptive threshold strategy; and finally, according to the trained network model, face attribute recognition results of the subjective attribute and objective attribute sub-networks are taken as final face attribute recognition results. A dynamic weight scheme and self-adaptive threshold adjustment are used, so that the network convergence speed is increased, and meanwhile, the label imbalance problem can be relieved; three different sub-networks are trained by adopting a spatial pyramid pooling method, so that end-to-end training is achieved to perform multi-task multi-face attribute recognition. And the precision of face attribute recognition, especially subjective attributes with high difficulty, is improved.

The invention provides a GRU based recurrent neural network multi-label learning method, which comprises the steps of S1, initializing a system parameter [theta]=(W, U, B); S2, inputting a sample {xi,yi}<i=1><N>, calculating a hidden state hT of the output of an RNN (Recurrent Neural Network) at each moment, wherein the sample xi belongs to R<M*1>, yi is a multi-label vector of the sample xi, andyi belongs to R<C*1>; S3, calculating a context vector hT and output zi of an output layer; S4, calculating the predicted output yi^, calculating the loss Li, and determining an objective function J;S5, solving the gradient of the system parameter [theta]=(W, U, B) according to a gradient descent method and a BPTT (Back-propagation Through Time) algorithm; S6, determining a learning rate [eta],and updating each weight gradient W=W-[eta]*[delta]W; S7, judging whether the neural network reaches stable or not, if so, executing a step S8, if not, returning the step S2, and iteratively updatingmodel parameters; and S8, outputting an optimization model. According to the invention, effective feature representation of the sample can be obtained by fully utilizing the RNN so as to improve the accuracy of multi-label classification. In addition, a problem of gradient disappearance is not easy to occur in back propagation.

The present invention discloses a feature selection based multi-example multi-tag learning method and system. The method comprises: mapping all features of a packet in a known data set into a projection example reference space consisting of all examples in the known data set, and obtaining a feature vector of the packet; removing projection examples corresponding to the features of the packet with an invalidly marked tag in the projection example reference space by adopting a feature selection method based on l2 and 1 norm constraints, and further obtaining a representative projection example reference space; mapping the features of the packet into the representative projection example reference space, and obtaining a new feature vector of the packet; and constructing a linear decision function according to the new feature vector of the packet, and training a classifier based on tag correlation by adopting an optimal algorithm based on the tag correlation. By using the method provided by the present invention, the classifier based on the tag correlation is obtained by learning by utilizing the known database, so that a tag set of unknown samples is predicted and the accuracy of tag prediction is improved.

The invention discloses a multi-label age estimation method based on a convolution neural network, which comprises the following steps: an input sample data set is obtained; face detection is carriedout on each input sample, alignment is carried out, and normalization is carried out according to the face position; the age tags of the input samples are subjected to multi-tag processing so that each sample is mapped to the same number of tags. all the normalized images are used as the input of the convolution neural network, and the multi-tag set is used as the output to train the convolution neural network, and the age estimation model is obtained; according to the principle of binary classification output and multi-label processing, combined with the ordered characteristics of age, the age of human face estimation is calculated. The invention utilizes the micro-variability and the orderliness between age features, constructs a convolution neural network model by using the idea of multi-label learning, and solves the problems of low age estimation accuracy and poor robustness existing in the existing age estimation method.

The invention provides a reinforcement patch and multi-tag learning combination-based expression lie test method. The main contents of the method comprise training of a lie test system and a lie test process of an expression. The process of the method comprises the steps of firstly importing a video to a database, and performing analysis and lie test on the expression so as to obtain a micro-expression lie test system; and when the lie test of the expression is started, inputting video data firstly, then performing expression identification, performing a query in the micro-expression lie test system, and finally returning a lie test result. According to the method, a lie test process video is shot by adopting a hidden camera, so that the psychology of subjects is not influenced and the authenticity of a test result is high; meanwhile, additional equipment attached to subject bodies is not required, and few manual operations are carried out in the whole lie test process, so that the loss of manpower and material resources is reduced; and the human face expression is intelligently analyzed without artificial participation, so that the result is more objective.

The invention belongs to the field of intelligent cell biological recognition, and relates to an RNA binding protein recognition method based on multi-view depth features and multi-label learning. Themethod comprises a training stage and a use stage, wherein the training stage comprises initial multi-view data construction, a depth multi-view feature extraction model and multi-label classifier training, wherein the initial multi-view data construction comprises the following steps: converting an original RNA sequence into an amino acid sequence and a dipeptide component by using a molecular biology principle and a statistics principle to obtain the characteristics of the amino acid sequence and the dipeptide component, and then constructing an initial multi-view characteristic together with the original RNA sequence, and constructing a model for the initial multi-view characteristic. According to the method, based on the initial multi-view data, the CNN is used for deep learning to construct a deep multi-view characteristic, and compared with the original multi-view characteristic, the multi-view characteristic extracted based on the deep features have smaller data dimensions anda higher classification effect.

The invention discloses a multi-label classification method based on flow pattern matrix completion, and the method gives consideration to an inter-feature flow pattern hypothesis in a process of multi-label learning carried out through employing matrix completion, and can improve the multi-label classification algorithm recognition efficiency more remarkably. Compared with the prior art, the method provided by the invention is greatly improved at three multi-label learning indexes (hamming loss, mean precision and coverage).

The present invention relates to an image labeling method based on multi-label learning, a terminal device and a storage medium. The method comprises the following steps of: S100: extracting all the labels in examples, and calculating the label importance degree of each label corresponding to each example in a multi-label training set through a label propagation method; S200: performing resamplingof the multi-label training set according to the label importance degrees to obtain training subsets; and S300: calculating class attributes of the training subsets, and performing classification according to the class attributes. Aiming at the problems of ambiguity and amounts of images in the image labeling field, multi-label learning is employed to perform labeling of the images, the information of the relative label importance degrees hidden in the training samples is employed to construct more effective class attributes, the image labeling method is employed to construct more effective class attributes corresponding to different labels and construct a classification model of the labels on the class attributes so as to obtain better image classification effects.

The invention discloses an automatic market analysis method based on multi-mode learning. The automatic market analysis method comprises the steps that firstly, a learning device is trained, and then in practical use, the trained learning device is used for predicting a market. A method for training the learning device comprises the steps that firstly, different-mode information of the market is collected and labeled; then, a multi-example generating method of market data features and character features is used for converting low-level features into a multi-example-packet format; finally, a multi-example multi-label learning method capable of utilizing various modes is adopted to conduct fusion processing on the data, and multi-label learning is conducted. According to the automatic market analysis method based on multi-mode learning, by acquiring multiple pieces of side information of the market, market changes can be depicted more comprehensively and market changes can be predicted more accurately. According to the method for conducting market prediction by using the multi-mode information, the different-mode data information can be used in the implementation process, adaptability is high and effects are good.

The invention discloses a post-physical examination chronic disease prognosis system based on multi-label learning. The system comprises a data acquisition module, a data preprocessing module, a basicprediction model construction module and a local prediction module. The data acquisition module is used for acquiring physical examination data of a physical examination user; the basic prediction model construction module is used for constructing a multi-label learning model for a physical examination scene; the local prediction module comprises a local model training unit and a prediction unit,the local model training unit solidifies the trained local prediction model into the local prediction module, the prediction unit outputs prediction prognosis indexes for a plurality of chronic disease occurrence conditions, and finally the future expected occurrence time of the chronic disease is obtained. According to the system, a multi-label learning method is used, the internal relation under the chronic disease concurrency condition can be extracted, the characteristic of high concurrency of chronic diseases is better met, and accurate prediction of the chronic disease occurrence condition in the future can be better completed.

The invention discloses a heterogeneous data generic feature learning and multi-label learning method and system, and relates to the technical field of language teaching. The invention aims to providea set of intelligent foreign language learning method and system so as to reduce the learning difficulty and improve the foreign language learning efficiency. The invention provides an intelligent foreign language learning method and system. When a foreign language article is displayed, new words in the article are automatically recognized according to a mature word list corresponding to a user,and the new words are highlighted and paraphrases and/or phonetic symbols and the like of the new words are displayed. When the foreign language article is read, selectable reading modes including a conventional reading mode, a short sentence reading mode, a word reading mode, a translation reading mode and a combined reading mode are adopted. The sequence and/or reading times of each reading modeincluded in the combined reading mode are/is set by the system or customized by a user.

The invention discloses a financial abnormity detection method and device based on multi-label learning. The method comprises the steps: generating a feature vector sample of each enterprise accordingto the financial information of an existing enterprise, carrying out the sample balance according to a preset sampling parameter, and obtaining a training sample set; labeling the training sample setaccording to a plurality of labels obtained according to historical responsibility asking data, and constructing a financial abnormity detection model based on a multi-label learning algorithm; and finally, obtaining financial information of a to-be-detected enterprise, constructing a sample input vector, and inputting the sample input vector into the financial abnormity detection model to obtaina detection result. By adopting the technical scheme of the invention, the problem of low accuracy caused by few data samples and incapability of detecting financial whitewash means can be solved, and the detection accuracy is improved.

The invention discloses an image data multi-label classification method. The method comprises the following steps: firstly, constructing a topological relation of pictures; then, using a multi-label propagation algorithm (CMLP) based on cooperation for solving the problem of semi-supervised multi-label learning (SSML), thereby obtaining a confidence matrix of unmarked pictures. According to the invention, the multi-label classification prediction performance of the pictures can be improved by fully utilizing the correlation among the plurality of labels of the multi-label image data, so that the accuracy and robustness are improved; the distribution information of the non-supervised image data can be mined, so that the method well solves the problem of semi-supervised multi-label learning,and is superior to the existing semi-supervised image data multi-label classification method in performance.

The invention discloses a semi-supervised multi-mark distance measurement learning method fusing local measurement, and the method comprises the steps: extracting training data of any multi-mark application scene, and marking a small amount of data; for labeled data, filtering out samples with the labeling occupancy smaller than a set threshold value, removing abnormal points of unlabeled data through clustering and other operations, and improving the sample quality; expressing the distance metric to be learned as a combined distance metric form; constructing an optimization item based on thedistance metric representation; constructing a multi-mark loss item and manifold regularization item joint optimization model, and learning distance measurement; mapping the original data to a distance measurement space, and then performing learning by using an existing semi-supervised multi-mark learning algorithm to obtain a semi-supervised multi-mark classifier fused with local measurement; andinputting a to-be-predicted sample into the classifier to obtain a labeled sample. By adopting the method, the manual labeling cost can be reduced, and the practical application of the multi-label learning framework is promoted.

The invention provides a method for diagnosing a bearing compound fault in an unknown target domain under a variable working condition. The method comprises the following steps: constructing a fault feature extractor consisting of a deep residual network based on a multi-layer domain adaptive method; inputting a preprocessed bearing vibration signal, and carrying out distribution difference matching on features, extracted through a plurality of residual blocks, of source domain data and target domain data to obtain migratable features; representing the composite fault as a combination of single faults through multi-label learning; and a binary association strategy is used to train corresponding binary classifiers for various single faults, and the features of the single faults are separated from the composite faults and are diagnosed respectively. According to the method, the problems that a traditional diagnosis scheme depends on expert knowledge and is difficult to effectively decouple and recognize the composite fault are solved, accurate diagnosis of the composite fault of the bearing under variable working conditions is achieved, meanwhile, dependence of an existing method on marked data is eliminated, and accurate diagnosis can be conducted on a related but invisible target domain.

The invention relates to a multi-label learning technology in the field of machine learning, and relates to a multi-label image deep learning classification method and equipment. The method comprisesthe following steps: acquiring a label relation graph; acquiring mapping of all types of labels and mapping of all label groups according to the label relation graph; constructing a deep convolutionalneural network and carrying out image general feature extraction; selecting feature maps of different layers of the convolutional neural network, and mapping the feature maps to a label and label group mapping dimension through a mapping function; calculating a conformity score and a normalization score of the label and the label group at the current pixel point position for all the pixel pointsin the selected feature maps; acquiring a final label related semantic feature and a final label group related semantic feature; and performing label prediction. According to the method, the label relationship is effectively utilized, richer image general features and label relationship features are learned, and a multi-label classification task is better carried out.

The invention belongs to the technical field of computer text classification, and particularly relates to a multi-label text classification calculation method based on ensemble learning, which comprises the following steps: step 1, preprocessing an original data set, segmenting sentences into independent words, and deleting non-keywords; : step 2, performing feature extraction vectorization processing on the text in a word frequency-inverse text frequency mode; : step 3, decomposing the multi-label learning problem into a plurality of independent binary classification problems by adopting a binary association method, wherein each binary classification problem corresponds to one label in the label space; and : step 4, classifying the labels by adopting an ensemble learning algorithm. The time complexity is reduced, the training speed is improved, the generalization ability of the weak learner is improved, the risk of over-fitting is reduced, and the robustness of the model is improved.

The invention provides a multi-label classification method and system, a readable storage medium and computer equipment, and the method comprises the steps: converting a multi-label classification problem into two types of classification problems, and enabling each two types of classifiers to correspond to one label in a multi-label data set; Constructing a semi-supervised multi-mark learning model according to preset feature selection and the correlation between the two types of classifiers; Solving the semi-supervised multi-label learning model by adopting a preset algorithm to obtain a semi-supervised multi-label classification model parameter and a label correlation matrix; And predicting a mark set to which the unknown mark belongs according to the model parameters and the mark correlation matrix. In the multi-label classification method, a semi-supervised multi-label learning method combining label correlation and feature selection is adopted, a large number of unlabeled multi-label samples are effectively utilized, correlation between labels is automatically obtained through learning, in addition, dimensionality reduction is conducted on high-dimensional data, and a multi-label classifier with the better generalization performance can be obtained.