86 results about "Naive bayesian classifier" patented technology

The invention discloses a real time filtrating method for large-scale garbage message based on the content, including the steps as following: 1, pre-filtrating by using the black list and the white list; 2, carrying out the online filtrating by using the filtrating module based on the frequency; 3, carrying out the fast filtrating for the message content by using the method of twice hashing; 4, carrying out the pretreating of the message text for suspicion message, and converting the same into the phase vector; 5, judging the suspicion message by using the method of combination of Naive Bayesian classifier and support vector classifier. The invention can greatly improve the filtrating speed of garbage message, and efficiently reduce the produced erroneous judgement rate in the conventional key word filtrating method; can efficiently solve the problem of group sending garbage messages with malicious intent in the short time; can efficiently avoid to mistake the common message as the garbage message so as to reduce the erroneous judgement, and efficiently improve the filtrating accuracy of whole system by analyzing the message content on the semantics.

The invention discloses a semi-supervised learning-based recommendation system shilling attack detection method. The method comprises the following stages: a shilling attack detection preprocessing stage, namely preprocessing the data of a marked data set and an unmarked data set to obtain shilling attack detection indexes of the marked and unmarked data sets, and training an initial Naive Bayes classifier on the marked data set; and an EM-lambda algorithm stage, namely continuously iterating by an EM-lambda algorithm to obtain a stable classifier and finally obtaining the type of the unmarked data set, and predicting whether an unmarked user belongs to a normal user (N) or a shilling user (S) through a function f: U to C so as to finish the recommendation system shilling attack detection. The method is used for discovering shilling attack users in a recommendation system, has high efficiency, sensitivity, special effect performance, high detection rate and low error rate.

The invention discloses a video object tracking method based on a self-adaptive measurement matrix. The method comprises the first step of selecting a tracking region, the second step of establishing the self-adaptive measurement matrix required by a compressed sensing sample according to the selected tracking region, the third step of calculating parameters corresponding to all nonzero elements in the self-adaptive measurement matrix, the fourth step of establishing a naive Bayesian classifier H as a target classifier for a target to be tracked, and initializing the target classifier, the fifth step of inputting a new video frame, evaluating compression measurement vectors of candidate targets according to the target classifier and then determining a tracking result, the sixth step of updating the target classifier according to the tracking result, and the seventh step of judging whether tracking needs to be continued according to whether a subsequent video frame exists. By means of the method, a more accurate and quicker tracking result can be obtained.

The invention discloses a logistics recommendation method based on clustering and cosine similarity. Firstly, the AP clustering method, the SDbw clustering method and the K-means clustering method are used to calculate the best K value for using the K-means clustering method of the cargo data set and the truck data set. According to the best K value, the cargo data set and the truck data set are subjected to clustering, and the naive Bayesian classifier is used to train two classifiers based on the results obtained from clustering of the cargo data set and the truck data set. The classifiers obtained by training of the truck data set and the cargo data set are used for classification, and then the cosine distances between normalized truck information and all elements of the same type of truck information in the truck data set are calculated, and finally the goods are recommended according to the cosine distances from large to small. The method effectively improves the real-time response speed of the recommendation method.

The invention discloses a fluorescent magnetic powder-based intelligent defect identification system. The system mainly comprises scene depth information collection, magnetic trace image preprocessing, color image segmentation based on a support vector machine (SVM) algorithm and a fuzzy C-mean (FCM) clustering algorithm, morphological processing and feature extraction, and crack identification based on a naive Bayesian classifier (NBC).

An electroencephalogram (EEG) feature extraction method based on dominant electrode combination and empirical mode decomposition (EMD) includes inputting n-channel EEG signal data, selecting dominant electrodes according to the principle that electrodes with record of EGG signals having classification performance higher than a certain threshold are called the dominant electrodes, otherwise, are called non-dominant electrodes, selecting dominant combinations, respectively extracting feature of EGG data of a training sample set and EGG data of a testing sample set corresponding to each dominant combination by means of the EMD to acquire training feature vector and testing feature vector of each dominant combination, respectively inputting the training feature vector, the testing feature vector, training sample set label, testing sample set label of each dominant combination input a naive Bayesian classifier to classify to obtain classification accuracy of each dominant combination, and finally surmising relation between different stimulations and activated brain regions when executing relevant motor imagery tasks according to the classification accuracy of each dominant combination.

The invention provides a chemical process fault monitoring method based on active learning, comprising: acquiring attributes of chemical process signals in real time; establishing an initial training signal set and an unlabeled observation signal set according to historical signals of a chemical process; updating the training signal set and the unlabeled observation signal set; establishing a naive Bayes classifier model according to the training signal set and forecasting class labels of unlabeled observation signals; correcting the naive Bayes classifier model by using an active learning method to obtain a final naive Bayes classifier model; and monitoring chemical process faults by using the real-time acquired attributes of the chemical process signals as inputs of the final naive Bayes classifier model. According to the method, a naive Bayes classifier is established on the basis of a finite number of labeled data and a large number of unlabeled observation data, and useful samples in the unlabeled observation signals are searched by using an active learning method to retrain the naive Bayes classifier, so that new fault types can be discovered and the classification precision can be improved.

The invention relates to artificial intelligence, and provides a model-based task allocation method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining agent feature information corresponding to a specified agent which is a agent in a to-be-allocated task state at present; inputting the agent feature information into a preset prediction model, generating a prediction task type corresponding to the agent feature information, wherein the prediction model is generated after a preset naive Bayes classifier is trained according to pre-collected sampledata; judging whether a to-be-processed task corresponding to the predicted task type exists in a preset task pool or not; if yes, screening out a specified task from the to-be-processed tasks according to a preset rule; and allocating the specified task to the specified agent. According to the application, the allocation efficiency and allocation accuracy of task scheduling of the agent can be effectively improved. The invention further relates to the field of block chains, and the prediction model can be stored in the block chain.

The invention provides a Chinese formal text word segmentation method based on active learning. The method comprises the steps that a current annotation data set L is used to train a naive bayes classifier; the current naive bayes classifier is used to annotate a to-be-annotated data set U; a sampling method is used to select a most informative fragment to be annotated for an expert; the new-sampled annotated fragment is added into the annotation data set L; and constant iteration is carried out until a preset satisfaction condition stops. The Chinese formal text word segmentation method basedon the active learning can effectively reduce artificial annotation data and obtain a tokenizer with better performance. The performance (measured by adopting an F value) of a model obtained by dataextraction and training by using an active learning method is about 5 percentage points higher than that of a model obtained by the data extraction and training by adopting a random drawing method. The performance each time of the model obtained by data extraction and training after the active learning is combined with EM iteration is improved by about 1.5 percentage points than that of the modelobtained by the data extraction and training by separately adopting the active learning method.

A method for separating out a defect image from a thermogram sequence based on weighted naive Bayesian classifier and dynamic multi-objective optimization. A method extracts these features and classifies the selected TTRs into K categories based on their feature vectors through a weighted naive Bayesian classifier, which deeply digs the physical meanings contained in each TTR, makes the classification of TTRs more rational, and improves the accuracy of defect image's separation. Meanwhile, the multi-objective function does not only fully consider the similarities between the RTTR and other TTRs in the same category, but also considers the dissimilarities between the RTTR and the TTRs in other categories, thus the RTTR selected is more representative, which guarantees the accuracy of describing the defect outline. The initial TTR population approximate solution for multi-objective optimization is chosen according to the previous TTR populations, which makes the multi-objective optimization dynamic and reduces its time consumption.

The invention discloses a naive-Bayes-based cable aging state evaluation method and apparatus. The method comprises the steps: collecting related parameters of a to-be-evaluated cable; and inputting the related parameters of the cable into a pre-constructed naive Bayesian classifier, and evaluating the aging state of the cable; and the pre-constructed naive Bayesian classifier construction methodcomprises the following steps: acquiring related parameters of an existing cable and a cable aging state, and taking the related parameters and the cable aging state as training data; taking the related parameters of the cable in the training data as attribute variables of a naive Bayesian classification model, taking a cable aging state as a class variable, calculating a prior probability of eachattribute variable based on a training data set, and establishing a naive Bayesian classifier. According to the method, the aging state of the cable can be accurately evaluated at low calculation cost based on a small number of samples, and targeted guidance is provided for operation and maintenance of the cable.

The method comprises the following steps: performing discretization preprocessing on data in a complete data set of a target stock in a target time period by adopting an equidistant discretization algorithm and a one-dimensional K-Means clustering discretization algorithm; carrying out attribute reduction of the technical indexes; adopting a naive Bayes classifier and a K-nearest neighbor classifier, and according to the complete data set subjected to attribute reduction, carrying out classification prediction on the increase and decrease amplitude of the target stock in the next trading day;and performing decision fusion on the classification prediction results of the future increase and decrease of the target stock obtained by the two classifiers by using a D-S evidence combination rule, and finally taking the decision fusion result as a final classification prediction result of the future increase and decrease of the target stock. According to the invention, the prediction accuracyof various stock trend prediction methods based on a neural network, an SVM and the like can be obviously improved. When the method is used for constructing a multi-factor stock selection model, thenonlinear relationship between various stock index data and stock income is more significant.

The invention provides a memory test method based on a naive Bayes classifier. The memory test method comprises the following steps: S1, selecting a memory introduction test item; s2, generating training data according to a production line yield corresponding to a test result of the memory introduction test item; s3, using the training data to train and classify the naive Bayes classifier; s4, carrying out production line yield prediction on the to-be-introduced memory by adopting the trained naive Bayes classifier; and S5, judging whether to introduce the to-be-introduced memory or not according to the predicted yield of the production line. According to the invention, a test item selection basis, a parameterization processing method and a naive Bayes and test item combination mode are adopted, an existing memory introduction test item is combined with a naive Bayes classifier artificial intelligence algorithm, the yield of a to-be-introduced memory production line is predicted, whether the to-be-introduced memory is good or bad is judged, and a controllable calculation support is provided for a component engineer.

The invention discloses an author disambiguation method based on incremental learning. The author disambiguation method comprises the following steps: obtaining a historical citation record, wherein the historical citation information has known clustering labels, and different clustering labels represent different author individuals; judging whether each clustering cluster is a clustering clusterof a first type or a clustering cluster of a second type according to the number of the historical citation records, and for the clustering clusters of the first type with a large number, training a corresponding naive Bayes classifier by using the feature vectors and clustering labels of the historical citation records; and screening out candidate clustering clusters, according to the types of all candidate clustering clusters, carrying out classification processing on the new citation records according to conditions, comprehensively using a naive Bayesian classifier to calculate the affiliated probability for classification, using the synergy person similarity to perform supplementary judgment on the affiliated probability mode classification, and calculating the semantic similarity withthe second type of clustering cluster to solve the problem that the naive Bayesian classifier cannot be used for probability classification. The author disambiguation method is good in author disambiguation effect and low in calculation overhead.

The invention discloses a defect prediction method for a numerical control system software module, which comprises the following steps of: 1) preprocessing a training data set, and determining parameters in a classifier; 2) performing data processing on the training data set, and dividing test data and training data; 3) establishing an improved naive Bayesian model based on a feature dependency relationship, processing features of the training data set in a pairwise manner, and performing model learning; 4) preprocessing the to-be-tested data by using the parameters in the classifier determined in the step 1); and 5) performing defect prediction on to-be-tested data by using the trained improved naive Bayesian model based on the feature dependency relationship. The dependency relationship is quantized and then added into the naive Bayesian classifier, so that the improved naive Bayesian classifier can be effectively applied to the numerical control system, and the purpose of accurately and efficiently detecting whether defects exist in the software module of the numerical control system can be achieved.

The invention relates to the field of geological drilling engineering, and provides a drilling process fault diagnosis method which comprises the following steps: acquiring a drilling process data setthrough a drilling machine sensor, and preprocessing the drilling process data set to obtain a preprocessed drilling process data set; selecting and combining drilling process parameters of the preprocessed drilling process data set through a distance correlation analysis method to obtain k parameter combinations; performing time sequence feature extraction on each parameter combination accordingto Mann-Kendall trend test and approximate entropy to obtain a feature combination corresponding to each parameter combination; clustering the feature combinations through a density peak search clustering algorithm to obtain clustering results of the parameter combinations; and according to the clustering result of each parameter combination, performing fault diagnosis through a naive Bayesian classifier to obtain a drilling process fault diagnosis result. Fault diagnosis in the drilling process can be effectively achieved, the safety of the drilling process is guaranteed, and economic and environmental losses caused by faults are reduced.

The invention relates to a facial expression recognition method and system. The method comprises the steps of identifying a face area of an image; extracting 68-point key points of each human face byutilizing Dlib to extract geometrical characteristics of the human face; wherein the geometrical characteristics comprise mouth height, mouth width, eye height, eyebrow height sum, eyebrow width sum,mouth minimum enclosing rectangle area, distance between an eyebrow and an eye head, eyebrow distance, lip bead and mouth corner height, and minimum enclosing rectangle area of two eyes; sending the geometrical characteristics of the human face into a naive Bayesian classifier, and determining an expression category; and during Bayesian decision, K neighbors are used for secondary learning to improve the classification effect. The scheme of the invention overcomes the defects of low classification efficiency, high calculation cost and influence on the effect of a classification algorithm caused by high dimension of existing face features.

The invention provides a festival sentiment analysis method based on naive Bayesian classification. Based on a natural language processing technology, a microblog influence model is established, and weight distribution is performed on comments, forwarding and like numbers of microblogs to obtain an influence value. And then starting from the specific festival direction, the naive Bayesian classifier is utilized to classify the oriented microblog data, so that the essential factors influencing the festival emotion acquisition of the user can be explored. And the extraction of the related factors also provides a theoretical basis for public opinion analysis and control and commercial marketing development.

The invention provides a naive Bayes lithofacies classification ensemble learning method and device based on characteristic randomness. The method comprises steps of multiple kinds of logging data ofa target work area being acquired and preprocessed; randomly sampling a plurality of preprocessed logging data into a training set and a test set according to a proportion; randomly generating a plurality of training subsets according to the feature combinations randomly selected from the training set and the component number of the feature combinations; training a plurality of first base classifiers in parallel by using a plurality of training subsets to obtain a plurality of second base classifiers and performance index values thereof; wherein the first base classifier is a naive Bayes classifier; determining a voting weight of each second base classifier according to the performance index value of each second base classifier; performing parallel lithofacies classification on the test set by utilizing a plurality of second base classifiers to obtain a classification sub-result of each second base classifier; and voting and combining the classification sub-results according to the voting weight to obtain a lithofacies classification result. According to the method, classification accuracy and learning efficiency of the lithofacies classifier based on naive Bayes can be improved.

The invention relates to a robot vision tracking method and system. The method includes: obtaining a trained naive Bayesian classifier by utilizing a Monte Carlo algorithm and a compressed sensing theory according to a frame image at the moment t; sampling the frame image at the t + 1 moment to obtain a plurality of images to be tracked; using a Monte Carlo algorithm and a compressed sensing theory to perform dimension reduction on the plurality of images to be tracked respectively to obtain a compressed feature vector of each image to be tracked at the t + 1 moment; and inputting the compressed feature vector of each to-be-tracked image at the t + 1 moment into the trained naive Bayesian classifier, and performing target and background classification to obtain a tracking target at the t +1 moment. According to the robot vision tracking method and system provided by the invention, the Monte Carlo and the compressed sensing are combined, so that the vision tracking efficiency and reliability are improved.

The invention relates to the technical field of aluminum electrolysis cell production, in particular to an aluminum electrolysis cell condition health degree classification method based on combined weighted naive Bayes. According to the method, common types, reasons and phenomena of disease cells are analyzed, an aluminum electrolysis cell condition health degree evaluation index system is established through an aluminum electrolysis mechanism, and then weighting is performed on a naive Bayes classifier in a combined weighting mode of an analytic hierarchy process and an entropy weight method. An aluminum electrolysis cell condition health classification model based on the combined empowerment Bayesian classifier is established, refined, quick and automatic recognition of the health degree of the electrolysis cell can be achieved, and the problems of subjectivity and retardancy existing in traditional manual judgment are solved. The weight is obtained by combining two modes, so that the naive Bayes attribute condition independence hypothesis is weakened, the real difference between indexes is enhanced, the problem of large error of numerical calculation weight when the sample size is small or missing is solved, and the weight is more accurate and has more practical significance.

The invention relates to a method and system for selecting a least-loaded route based on a naive Bayes classifier, so that the performance of a method for selecting a least-loaded route is improved. A network snapshot records historical network status information, and a naive Bayes classifier is used to predict the potential future network blocking probability if a service connection is established along a candidate route between each node pair. A network snapshot corresponds to each service request that arrives, and records the number of busy capacity units on each link.

The invention provides a disease inference system based on a knowledge graph and a self-adaptive mechanism, and relates to the technical field of computer technologies and auxiliary medical machinery. A disease inference model is constructed by taking the knowledge graph of triad as a data structure and taking a TransE translation model and a naive Bayes classifier as a basis, and the disease inference model is used for inference. Firstly, a data set is divided, a proper number of TransE translation models and naive Bayesian classifiers are searched for each part of the data set to be matched, self-adaptive matching is achieved, the influence of input imbalance on the models can be reduced, the interpretability can be improved through naive Bayesian classifiers, the disease reasoning accuracy of a system is improved, the disease reasoning accuracy of the system is improved, and the working efficiency of medical workers is further improved.

The invention relates to a dual-polarization radar secondary echo recognition method based on a naive Bayes classifier, and the method comprises the steps: calculating four criteria, namely PDEh, PDEv, PDEhv and SQI, of a current distance library for secondary echo recognition through employing dual-polarization radar I/Q data; according to the numerical values of the four criteria, finding out each criterion CPDF which is counted in advance, obtaining the probabilities of the specified numerical values under different classifications, and calculating the probability Pf that the distance library echo data belong to primary echoes and the probability Ps that the distance library echo data belong to secondary echoes; when Ps is greater than Pf, determining that the distance library echo is a secondary echo, otherwise, determining that the distance library echo is a primary echo; and continuing to identify the next distance library. According to the method, the secondary echo recognition precision in the radar data level is improved, and the method has great significance in improving the dual-polarization radar parameter estimation quality.