104 results about "Naive Bayes classifier" patented technology

The invention discloses an electric power user figure establishment and analysis method based on the big data technology. The method comprises steps that the historical electricity information, basic attributes, the fee-paying information and the appeal information of electric power users are acquired; classification category sets of user figures are determined, an influence factor set of a classification result is determined, and a mapping relationship between the influence factor set and the classification set is determined; random extraction of the acquired data is carried out, one part of the data is taken as a training sample, and other data is taken as prediction sample; normalization processing, discretization processing and attribute reduction for the training sample and the prediction sample are carried out, and an influence factor set after correction is determined; the training sample is trained, ten-fold cross validation is taken as a test mode, an electric power user figure prediction model based on a naive Bayes classifier is established, data classification mining analysis on the prediction sample is carried out through utilizing the prediction model, and electric power user figures are acquired. The method is advantaged in that electric power electric quantity prediction and management can be facilitated.

The invention discloses a method and a system for detecting a Chinese phishing website. The method comprises the following steps: S1, acquiring a URL (Uniform Resource Locator) by a client; S2, extracting URL features and website page content features as feature vectors respectively; S3, performing classification trainings on the feature vectors through an SVM (Support Vector Machine), an extended website page content feature NBC (Naive Bayes Classifier), a decision tree algorithm and a link and form processing method; S4, performing classification integration on classification training results, and judging whether a website with the URL is a phishing website. According to the method and the system, the URL features and the website page content features are extracted as the feature vectors, the corresponding classification trainings are performed through the SVM, the NBC, the decision tree algorithm and the corresponding processing of links and forms, prediction results are integrated by virtue of classification integration to obtain a final result, and therefore the classification accuracy is greatly improved.

A method is provided for deriving optimized operating parameter settings for industrial furnaces of different designs as commonly used in power generation that will achieve robust and desirable operations (for example, low NOx and low CO emissions while maintaining specific furnace exit gas temperatures). The method includes the application of recursive partitioning algorithms to historical process data to identify critical combinations of ranges of operational parameter (combinations of settings) that will result in robust (low-variability) desirable (optimized) boiler performance, based on empirical evidence in the historical data. The method may include the application of various algorithms for recursive partitioning of data, as well as the consecutive application of recursive partitioning methods to prediction residuals of previous models (a methodology also known as boosting), as well as the application of other prediction algorithms that rely on the partitioning of data (support vector machines, naive Bayes classifiers, k-nearest neighbor methods).

The invention discloses a fault diagnosis method and device for a rolling bearing of a running gear of a locomotive. The method includes the steps of collecting vibration acceleration data of the rolling bearing of the running gear of the locomotive under different fault types, and grouping the data according to the fault types; obtaining frequency domain signals of the vibration acceleration data according to the grouped vibration acceleration data; conducting three-layer wavelet packet decomposition on the frequency domain signals, and constructing fault characteristic sets; randomly arranging the fault characteristic sets, using the front B sets as the training sets, and using the rear C sets as the testing sets, wherein the sum of B and C is equal to A, and B is larger than C; training the B training sets through a Naive Bayes classifier, and establishing fault diagnosis model based on Naive Bayes for the rolling bearing of the running gear of the locomotive; classifying the C testing sets according to the fault diagnosis model, and evaluating the classification performance of the fault diagnosis model according to the classification result and through the combination with the fault characteristic sets.

The invention discloses a compressive sensing-based real-time multi-scale target tracking method. A sample is modeled by extracting the normalized rectangle features of sampled image, and the normalized rectangle features have higher robustness for the multi-scale target tracking. The normalized rectangle features are very high in dimensionality, so that the method can be used for compressing high-dimensional features based on compressive sensing, the feature vector is compressed under the condition that the extraction scale is not changed, the computation complexity is greatly reduced by integrogram, and the demand of real-time tracking can be met. The compressed feature vector of the sample is classified by a Naive Bayes classifier, so that the most probable position of a target can be determined; the classifier is used for responding and estimating the particle weight and resampling particles so as to prevent the degeneration of particle tracking capability; furthermore, a second-order model is used for estimating and predicting the particle state under the condition that the target movement speed factor is considered. The target in video image can be tracked in real time by the compressive sensing-based real-time multi-scale target tracking method; the method is high in accuracy and low in computation complexity; a tracking frame changes in real time along with the change of target scale, so that the demand of actual tracking application can be met.

The invention discloses a software service quality monitoring method and system based on a weighted naive Bayes classifier. By the constructed weighted naive Bayes classifier, the QoS (quality of service) of software is judged to determine that the QoS belongs to a standard satisfying category, a standard unsatisfying category or an incapable of judging category. During training, an influence factor combination is set, influence factors refer to software's influence on QoS, and the weight and the priori knowledge of the influence factor combination are calculated; during monitoring, the classifier acquires monitoring results, and analyzes, stores and returns evaluation to a data server. The system comprises a controller, an observer, a trainer, an optimization sample set and an analyzer, wherein the controller collects different QoS statements, issues data standard instructions required by different QoS, transmits QoS standards which need to be matched with data sets to the trainer, and controls collecting cycle and frequency; a database collects the monitoring results of the analyzer; a service capability evaluating module returns software monitoring results and evaluation results to data service equipment.

The invention belongs to the field of proteomics classification, relating to a classification method of differential proteomics. The method comprises: selecting characteristics by univariate statistics analysis, sequential feature selection and a genetic algorithm; extracting characteristics by main ingredient analysis and a partial least squares method; connecting and integrating with linear discriminant analysis, a k-nearest neighbor classifier, a support vector machine, a decision tree, a naive Bayes classifier and an artificial neural network in series to obtain serial integrated classifiers which are connected in parallel and combined; endowing each base classifier with a weighting coefficient according to classification accuracy rate; and taking a fuzzy attribute value as a classification result output mode to obtain the classification judgment result of a target sample. The method has higher classification accuracy rate and robustness on classifying and judging abnormal and normal samples in the differential proteomics, and is suitable for classifying and analyzing the data of differential proteomics.

The invention discloses a network failure diagnosis method based on a selective hidden Naive Bayesian classifier, comprising: (1), obtaining history data from a network history database, wherein the history data comprise a symptom variable set and a failure class variable set; (2), constructing a selective hidden Naive Bayesian classifier prediction model, determining corresponding most related symptom variable set according to every symptom variable in the symptom variable set; (3), automatically learning classifier parameters by the selective hidden Naive Bayesian classifier through training the history data; (4), in failure diagnosis, estimating the test data by using the selective hidden Naive Bayesian classifier so as to obtain corresponding final failure diagnosis result. Through executing the network failure diagnosis method of the invention, the problems in the existing network failure diagnosis that the operation complexity is high and the network diagnosis result is great in deviation are effectively solved; the network diagnosis accuracy is greatly improved; the operation complexity is further reduced, and better learning capability and fault-tolerant character are kept at the same time.

The invention provides an Android malware detection method based on an improved Bayesian algorithm. The feature attributes of Android malicious programs and well-behaved programs are analyzed and classified through the improved Bayesian algorithm to realize the malware detection method based on the improved Bayesian algorithm. A judgment on whether software is malware is implemented from the aspect of permission application of applications. According to the method, a permission request label in an Android permission request mechanism is taken as a detection data source. The malware and well-behaved software are distinguished in a permission request label combination way, and a detection model is built by using the improved Bayesian algorithm. The improved Bayesian algorithm is characterized in that mutual independence among attributes of the data source is considered, and a naive Bayesian classifier is used for performing data modeling, so that the detection index is increased greatly, the detection accuracy is increased, and the false alarm rate is lowered.

The invention provides a power transmission line fog level recognition method and system based on images. The method includes the steps that a plurality of training images of a power transmission line in the weather of sunshine, light fog, fog, heavy fog, smog and thick smog are collected; image classes corresponding to the sunshine, the light fog, the fog, the heavy fog, the smog and the thick smog are established according to the training images; characteristics of the image classes corresponding to the sunshine, the light fog, the fog, the heavy fog, the smog and the thick smog are extracted respectively; the characteristics of the image classes are used as input data of a naive bayes classifier to be trained so as to obtain a fog level recognition template base; an image to be recognized of the power transmission line is collected; characteristics corresponding to the image to be recognized are extracted; recognition is conducted on the characteristics corresponding to the image to be recognized according to the fog level recognition template base to obtain a recognition result; the recognition result of the image to be recognized is output. The fog levels are divided into the light fog, the fog, the heavy fog, the smog and the thick smog according to horizontal visibility distances, and classification recognition of the fog levels is achieved.

The invention provides a Bayes classification-based method for fusing traditional meteorological data with perception data. On the basis of a Naive Bayes classifier, the invention discloses a One-Dependence Attribute Weighted Naive Bayes method, to improve a conventional Naive Bayes algorithm, appropriately release the limit that attributes need to be independent from each other, and find a compromising point between the efficiency and the classification efficiency, so as to accomplish the fusion of radar data with user perception data. The method comprises the following steps of: preprocessing the data; constructing the classifier according to training sample data; and classifying samples to be classified by using the constructed classifier.

The invention provides a video tracking method based on local background learning. The video tracking method includes the steps that the time-space relationship between a target to be tracked and the local background of the target is modeled through the Bayes frame, a plurality of multi-dimensional images of the target are simultaneously collected through the time-space relationship between the modeled target and the local background, and dimensions of the collected multi-dimensional images of the target are reduced through a random sensing matrix meeting compressed sensing conditions to obtain feature vectors of the multiple multi-dimensional images; according to the feature vectors of the multiple multi-dimensional images, the multi-dimensional images with the dimensions reduced are classified through a naive Bayes classifier, and the position where the target appears is estimated according to a likelihood confidence image of the target position; based on target structure constraint conditions, a collector outputs the target with the maximum degree of overlapping with the previous frame target tracked successfully as the final tracking target. The video tracking method is suitable for video target tracking under complex conditions, and is high in discernment capacity and tracking accuracy.

The invention discloses a calculation method of crime degree of speech data, and belongs to the technical field of intelligent security control. The invention provides a concept of the crime degree of speech. The concept of the crime degree of speech is defined as the crime possibility which is presented by a certain ID (Identity) on a social network through the speech of the ID, and an influence model of a demand factor, an emotion factor and a preparation factor of the crime degree of speech is proposed; and a naive Bayes classifier is applied to judge the demand factor by a text analysis technical means, the emotion factor is judged by an emotion dictionary, a crime sensitive word dictionary is constructed and is combined with a machine learning method to judge the preparation factor, and a crime degree theoretical frame of the speech and a mathematic model are established. The calculation method can cause early warning to be advanced to criminal psychology formation and criminal preparation stages, can automatically analyze and predict a great quantity of data in the whole process when the calculation method is applied to a practical network, does not need human intervention and can intelligently improve a security and protection system to a higher level.

The invention discloses a multi-target device-free localization method based on radio tomographic imaging. The method includes following steps: radio tomographic imaging; local maximum value extraction; cross section scanning: performing grayscale cross section scanning on each maximum value point after obtaining all local maximum value points in an imaging result, regarding the maximum value point as the center of a scanning line L, rotating the line with a fixed angle spacing theta, recording pixel values of pixel points which are penetrated by the line after each rotation, and stopping rotation after rotation of a lap; feature calculation of a grayscale distribution map; and classifying to-be-classified points into three categories by employing a naive Bayes classifier according to features of the grayscale distribution map: phony-target points, single-target points, and double-target points, determining each hotspot, removing phony targets, and distinguishing the target number when a plurality of targets are collected together. According to the method, the precision is high.

The invention discloses a video target tracking method based on dynamic sparse projection. The method is used for solving the technical problem that an existing fixed sparse projection matrix tracking method is poor in robustness. According to the technical scheme, different low-dimensional image feature information is obtained from a high-dimensional image by utilizing a series of sparse projection matrixes with different dimensions, and corresponding classified samples are respectively obtained through a naive Bayes classifier on this basis; the weight information of each classified sample is obtained by calculating the character contrast ratio of each classified sample and a previous frame of sample, the image similarity degree of each classified sample and the initial frame of sample and the comparison result of the pixel distribution difference degree of the current frame of target and a background, the sparse projection matrix with a weight value smaller than a threshold value is dynamically updated, and the classified sample with optimal weight is selected as the final target tracking result. The accuracy rate of the tracking result reaches more than 85%.

The invention discloses a student browsed webpage classification method based on N-Gram and a naive Bayesian classifier. The method comprises the specific implementation steps that first, URL description information is crawled from a navigation website, four classification corpora are constructed, corpus texts are expressed in the forms of uni-gram and bi-gram, TF-IDF is used as a weight of text characteristics, and a naive Bayesian classification algorithm is used to construct the classifier; and URLs in student browsed records are segmented according to set rules, URL categories are determined through matching of the classifier and a URL category base, and if the URL categories determined through the classifier conform to set confidence, the URL categories are added into the URL category base. Through the method, the URLs in the student browsed records are effectively classified, and therefore the webpage recognition rate and the classification accuracy rate are increased.

The invention discloses a road boundary point automatic extracting and vectorizing method based on on-vehicle laser scanning data. The method comprises the steps of a first step, calculating the characteristic of each laser footpoint in three-dimensional laser point cloud data; a second step, according to the characteristic of each laser footpoint, classifying the laser footpoints for obtaining road boundary points and non-road-boundary points by means of a naive Bayes classifier, and marking the obtained road boundary points as initial road boundary points; a third step, establishing a KD tree by means of all initial road boundary points, and respectively calculating the directional characteristic of each initial road boundary point; a fourth step, according to the directional characteristic of the initial road boundary point, clustering the initial road boundary points by means of the KD tree; and a fifth step, calculating the characteristic of each clustering area, eliminating the clustering areas which do not satisfy a preset condition, and obtaining a road boundary point extracting result. The road boundary point automatic extracting and vectorizing method improves automatic degree and production efficiency in point cloud data processing. Furthermore the road boundary point automatic extracting and vectorizing method has advantages of simple operation, easy realization and high practical value.

The invention discloses a shilling attack detection method based on a stack type sparse self-encoder. The shilling attack detection method is mainly used for solving the problem that corresponding characteristics of shilling attack users of different types need to be extracted in the prior art. The shilling attack detection method comprises the following steps: (1) inputting an initial scoring data set; (2) initializing the initial scoring data set; (3) directly using the score of each user as input to train the stack type sparse self-encoder, and extracting characteristic data of the user; (4) using the extracted characteristic data as input to train a Naive Bayes classifier; and (5) calculating the probability of a user with an unknown type belonging to each type according to the trained Naive Bayes classifier, and finding out the shilling attack user. The shilling attack detection method disclosed by the invention can be used for directly using the stack type sparse self-encoder to extract the characteristic data of each user, can be used for stably detecting shilling attack users of a plurality of types and can be used for detecting malicious attack users in an Internet system.

Disclosed is a water quality toxicity detection method based on fish activity analysis. The method comprises the following steps that 1, crucian is adopted as a biological monitoring object so as to be subjected to real-time monitoring; 2, a target crucian contour is extracted through conversion from RGB to HSV color space, crucian groups are monitored in real time, and a crucian tracking video sequence is obtained; 3, crucian motion data analysis and detection are performed, wherein 3.1, differences among the crucian velocity speed, a counter area and an area mean value are adopted as main characteristic data; 3.2, a mature detection model is generated on the basis of a Naive Bayes classifier algorithm; 3.3, novel characteristic data is adopted for detecting and judging whether the detection model is mature or not; 3.4, real-time water quality data is detected in an online mode through the mature detection model, and finally online detection of the water quality toxicity is achieved. Online real-time detection can be achieved, sensitivity and continuity of water quality detection can be improved, the detection cost can be lowered, and real-time effective detection can be performed on a large number of unknown water quality toxicity conditions.

The invention provides a sintering process working condition identification method and system considering a time sequence. The method takes the time sequence data of the process parameters of a sintering process as the input and takes a sintering process working condition as the output, and comprises the steps of firstly, utilizing a Spearman rank correlation analysis method and an information entropy analysis method for parameter selection and combination, and obtaining the combined decision parameters; then, using a fuzzy C-means clustering algorithm based on the dynamic time warping distance for clustering the time sequence data, and obtaining a clustering result of the combined decision parameters; and finally, using a naive Bayes classifier for the working condition recognition, and obtaining the recognized sintering process working condition. According to the working condition recognition method, the effective recognition of the working condition during the sintering process is achieved, and the important economic value and the application value are achieved.

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 target tracking method based on a dynamic measurement matrix and a target tracking system based on the dynamic measurement matrix. The method comprises the following steps of: 1, compressing high-dimension features of samples into low-dimension features, and initializing the dynamic measurement matrix; 2, collecting a plurality of positive sample sets and negative sample sets around a target position to perform classifier updating learning; 3, determining the position of a current frame target; and 4, updating the dynamic measurement matrix, and returning to the second step until the tracking is completed. The dynamic measurement matrix is used for extracting compression features of the target in the tracking process, i.e., in the tracking process, the measurement matrix is updated by utilizing the features of the tracked target and the features of a Naive Bayes classifier; the fixed form of the measurement matrix is improved; and the adaptability of the tracking method is high. Experimental results show that the tracking method has good robustness.

The invention discloses an intelligent evaluation and diagnosis method and a system for heart disease types and severity degrees. The method comprises the steps of acquiring disease characteristic data and demographic characteristic data, and analyzing the acquired ultrasonic echocardiogram report data and the patient demographic characteristic data by utilizing a learning model to obtain a modelevaluation index, a heart disease type and a heart disease severity. According to the invention, a data mining method is adopted, so that data preprocessing, data screening and other operations are carried out on data through the data mining correlation method. The method is adopted for selecting a noise ratio during the characteristics selection process. A random forest model is adopted for carrying out the classification prediction of the heart disease severity. Meanwhile, an effective research method is obtained through comparing and analyzing the algorithm performances and the learning effects of the random forest model, a naive Bayes classifier, a decision tree model and a BP neural network model. Moreover, a standard for the severity classification of heart disease patients and a prediction method for predicting the treatment risk of the heart disease operation are provided.

The invention provides a work order classification method and a work order classification device. The method comprises the steps of performing word segmentation on a work order to obtain a word set; acquiring a ratio of positive part of speech words and negative part of speech words in the word set according to a part of speech dictionary; if the ratio of the positive part of speech words and thenegative part of speech words satisfies a preset threshold value, determining that the work order is a work order category corresponding to the preset threshold value; and if the ratio does not satisfy the preset threshold value, judging the word set by use of a trained naive bayesian classifier to acquire the work order category to which the work order belongs. According to the method, preliminary classification is performed through the ratio of the positive part of speech words and the negative part of speech words in the word set so as to reduce the work order quantity needing to be classified by the naive bayesian classifier and improve the classification efficiency; and the method is suitable for classification of the mass work orders in the telecom industry and has relatively high accuracy.

The invention discloses a compressed irregular block LBP-based video target tracking method. The method comprises the steps of expressing a tracked target or a candidate target by adopting a compressed and sampled irregular block LBP eigenvector; searching for the candidate target by adopting a particle filtering framework; and judging whether the compressed eigenvector of the candidate target is a target tracking result or not by adopting a Navie Bayesian Classifier. According to the method, the processing speed of video target tracking is increased, and an accurate tracking effect can be kept in various complex scenes of severe illumination change and posture change, view angle rotation and sudden movement, background disorder, similar target interference and the like.

The invention discloses a method for identifying human activities based on a BP (Back Propagation) neural network in an intelligent family environment. The method for identifying human activities comprises the steps of: firstly labeling the data of various types of human activities, collected by a motion sensor and a project sensor in an intelligent family environment test board, and extracting the characteristics of the labeled data of the sensors; then inputting the extracted characteristic data to a BP neural network model by adopting a 3-fold cross validation method to be trained and identified; and finally comparing the identification result of the human activities based on the BP neural network with a hidden markov model method and a naive bayesian classifier method, wherein the computed result indicates that the identification accuracy is better by adopting the method for identifying human activities based on the BP neural network. According to the method for identifying human activities based on the BP neural network, the data is obtained by the sensors without the need of installing a video camera at the residence. Therefore, the method disclosed by the invention is easy to be accepted by residents, the data of the sensors is easier to process compared with the video data, the working amount is reduced, and privacy of the residents is protected.

The invention discloses an emotion dictionary construction method capable of being automatically updated and used for financial text analysis. The method comprises the following steps of: forming a basic dictionary Dinial by utilizing an existing sentiment dictionary in a knowledge base; the basic emotion dictionary is expanded by means of machine adding and manual adding; obtaining an extended emotion dictionary Dextend; improving the new word extraction accuracy by calculating the prefix and suffix information entropy, then conducting probability calculation on new words extracted from a corpus through a naive Bayes classifier and the emotional tendency probability, and adding the emotional words which meet the condition and have positive or negative emotions into an emotional dictionary by setting a threshold value. Compared with the prior art, the method has the advantages that (1) new words are extracted more accurately, and noise and subsequent calculation amount are reduced; (2) the emotion analysis calculation amount is small, and a more accurate emotion analysis result can be obtained through parameter optimization; and (3) the sentiment dictionary can be continuously updated as required, so that the accuracy of the financial text sentiment analysis method based on the sentiment dictionary is improved.

The invention relates to the machine learning technology field and particularly relates to an online frozen gait detection method. The method comprises steps that offline gait data and an offline gaitvideo in the offline walking process of a patient are obtained, the offline gait data includes thigh acceleration, thigh angular velocity, calf acceleration, calf angular velocity and plantar pressure; an offline sample set is established according to the offline gait data and the offline gait video; an offline normal gait and offline frozen gait naive Bayesian classifier is constructed; throughthe offline normal gait and offline frozen gait naive Bayesian classifier and online gait data, online normal gait and online frozen gait probabilities are obtained in the online walking process, andthe detection result is obtained. The method is advantaged in that the frozen gait can be rapidly and accurately detected in real time, and the patient is timely facilitated to recover to realize walking and normal movement.

The invention discloses a weighted naive Bayes indoor positioning method based on attribute independence, and belongs to the technical field of indoor positioning, and the method comprises the following steps: building a CSI sample set of a position point; performing CSI data preprocessing; extracting main features through a PCA algorithm; establishing an offline fingerprint database; in the online stage, using a weighted naive Bayes positioning algorithm with independent attributes; in the offline stage, through multiple times of sampling analysis, knowing that CSI amplitude values of any position obey normal distribution, and therefore the mean value and the variance of the amplitude values of all the positions serve as fingerprints to be stored. In the online stage, the variance contribution rate calculated in the principal component analysis stage is used as a weight to be applied to naive Bayes classification, and the advantages of principal component analysis are maximized. According to the method, only the mean value and the variance of the CSI amplitude values measured by each reference point for multiple times need to be selected as fingerprints, the data is processed by using the principal component analysis method, the conditional independence assumption of the naive Bayes classifier is met, and the positioning precision is improved.

The invention discloses a method for automatically classifying, obtaining and storing complex knowledge of a high-end device. The method comprises: an automatic complex knowledge classification method of performing induction and reorganization on knowledge resources from the following three dimensions of the high-end device: a life cycle dimension, a knowledge manifestation pattern dimension and a knowledge theme dimension, and automatically classifying the knowledge resources by using a naive Bayes classifier; a complex knowledge obtaining method of obtaining a template according to complex knowledge based on a meta-knowledge model and obtaining the complex knowledge resources through semi-automatic obtaining technology based on the obtained template; and an automatic complex knowledge storage method of dividing the complex knowledge resources from the physics through a series of automatic division rules, compressing key information and storing the same in different storage spaces in a distributed manner. The method disclosed by the invention covers the automatic complex knowledge classification method, the complex knowledge obtaining method and the automatic complex knowledge storage method, and provides foundation and support for the high-end device manufacturers to use the complex knowledge resources.