79 results about "Sub classification" patented technology

Techniques for managing big data include retrieval using per-subject dictionaries having multiple levels of sub-classification hierarchy within the subject. Entries may include subject-determining-power (SDP) scores that provide an indication of the descriptive power of the entry term with respect to the subject of the dictionary containing the term. The same term may have entries in multiple dictionaries with different SDP scores in each of the dictionaries. A retrieval request for one or more documents containing search terms descriptive of the one or more documents can be processed by identifying a set of candidate documents tagged with subjects, i.e., identifiers of per-subject dictionaries having entries corresponding to a search term, then using affinity values to adjust the aggregate score for the terms in the dictionaries. Documents are then selected for best match to the subject based on the adjusted scores. Alternatively, the adjustment may be performed after selecting the documents by re-ordering them according to adjusted scores.

Disclosed are methods of identifying subjects with Diabetes or a pre-diabetic condition, methods of identifying subjects at risk for developing Diabetes or a pre-diabetic condition, methods of differentially diagnosing diseases associated with Diabetes or a pre-diabetic condition from other diseases or within sub-classifications of Diabetes, methods of evaluating the risk of progression to Diabetes or a pre-diabetic condition in patients, methods of evaluating the effectiveness of treatments in subjects with Diabetes or a pre-diabetic condition, and methods of selecting therapies for treating Diabetes or a pre-diabetic condition, using biomarkers.

The invention relates to a method for classifying application programs of an embedded terminal. The main classification corresponds to a main list and each sub-classification corresponds to a sub-list; the records of all application programs are memorized in the main list and the records of the application programs of the sub-classification are memorized in the sub-list. The method comprises the following steps of A1) obtaining records corresponding to the application program icons from the main list; A2) searching the records corresponding to the application program icons in all sub-lists; A3) deleting the records from the sub-list containing the records; A4) adding the records to the sub-list corresponding to the first classification; and A5) displaying the icons corresponding to all records in the sub-list corresponding to the sub-classification in the main menu. The method for classifying application program of the embedded terminal carries out the management operations such as classification, movement, and the like, freely to the application icons, reinforces the independency of the user, leads the user to quickly and conveniently search the needed applications and can carry out the individuated naming and selection and display of individuated icons to the classified names of the shortcut icons of the application program.

Methods, data structures, and systems for generating customer segmentation models are provided. Basket transactions are analyzed and classified into a first segment type, a second segment type, a third segment type, or a fourth segment type. A number of the basket transaction within a number of the segment types are separately analyzed to determine sub classifications or sub segments within a particular segment type. Each basket transaction is augmented with a segment type that identifies the segment type classification, and a number of the basket transactions include a segment identifier that identifies the sub segment within a segment type that a basket transaction is associated with. The augmented basket transactions represent a customer segmentation model. In one embodiment, daily transactions are monitored by a script and used to dynamically adjust the customer segmentation model.

Disclosed are methods of identifying subjects with arteriovascular disease, subjects at risk for developing arteriovascular disease, methods of differentially diagnosing diseases associated with arteriovascular disease from other diseases or within sub-classifications of arteriovascular disease, methods of evaluating the risk of arteriovascular events in patients with arteriovascular disease, methods of evaluating the effectiveness of treatments in subjects with arteriovascular disease, and methods of selecting therapies for treating arteriovascular disease.

The invention discloses a method for early classifying imbalance multi-variable time sequence data. The method comprises the following steps: carrying out the under-sampling for a large category of category data sets according to an imbalance ratio for the inter-category data scale imbalance problem into a plurality of subsets, and combining the subsets with a small category of category data to form a plurality of training subsets; extracting and selecting core features of each training subset, and establishing a sub-classifier based on a rule according to the core features, wherein the feature selection process is realized in a clustering way in order to solve the imbalance problem of the data scale of the intra-category sub-concepts so as to guarantee the diversity of the core features; and solving the weight of a classifying effect of data in the training set by utilizing the sub-classifier on the basis of each sub-classifier, and establishing an integrated classifier. The classifier can solve the multi-variable time sequence classifying problem of the imbalance data set, the accuracy is relatively high, and the earliness degree is good.

Genes that are differentially expressed in subtypes of renal cell carcinomas are disclosed as are their polypeptide products. This information is utilized to produce nucleic acid and antibody probes and sets of such probes that are specific for these genes and their products. Methods employing these probes, including hybridization and immunological methods, are used to determine the subtype of a renal cell tumor sample from a subject based on the differential expression of such genes that is characteristic of the cancer subtype.

The invention provides a traffic mode behavior recognition model construction method. The method comprises the steps of 1) conducting the multi-level classification on specific traffic mode behaviors and establishing a corresponding classification tree; 2) for the classification problem of each parent node on the classification tree, selecting an optimal feature set for a test sample corresponding to the classification problem based on a random forest model according to the identification accuracy; 3) for the classification problem of each parent node on the classification tree, based on the corresponding optimal feature set obtained in the step 2), selecting one optimal classification model as a sub-classification model corresponding to the parent node. The invention also provides a corresponding traffic mode behavior recognition method. Based on the method, specific traffic mode behaviors can be distinguished more accurately. Meanwhile, the computational complexity is reduced.

The invention discloses a method and a device of file classification and generation of support vector machine model. The method comprises the steps of confirming classification of a file to be classified according to feature vectors of the file to be classified and the support vector machine model generated through a training set processed through classification flattening, wherein the classification flattening processing process of the training set comprises the steps of sequencing, according to levels of classifications from top to bottom, preset classification of each training sample in the training set; judging whether a sub-classification exists in the classification of each sample from a classification relatively high in level; and deleting the classification from the classification of the training sample if the sub-classification exists in the classification of each sample. Due to the fact that classification flattening treatment is conducted according to the level relationship between classifications, the obtained support vector machine model is enabled to be suitable for classification of multi-level classifications, and the classification is enabled to have better accuracy.

A mobile device such as a tablet computer accesses a server to search a product catalog. The catalog is displayed as a broad classifications of products, which when selected result in the display of sub-classifications within the classification. Levels of searching are provided to the product level, and movement within each level is possible.

The invention discloses a multi-label diagnostic method for train air conditioning faults. First a multi-label classification mechanism is established using proximity support vector machines as a basic classification algorithm, then actual normal operation data of a train air conditioning unit and various fault operation data are collected as training samples, a plurality of single-label sub-classifiers are obtained through learning and training, finally the actual operation data of the train air conditioning unit are used as inputs of the single-label sub-classifiers, a plurality of sub-classification result vectors are obtained, and the result is the solution of the multi-label classification problem. By adoption of the multi-label diagnostic method for train air conditioning faults, precision of fault diagnosis is improved, and a diagnosis result is relatively comprehensive.

A classification dictionary updating apparatus includes an update proposal receiving unit that receives a proposal for updating a hierarchical classification dictionary which has a hierarchical structure which includes a class that defines the hierarchical structure, a property that defines a hierarchical class structure, and an attribute that is a detailed information field of the class and the property, and in which a sub classification class inherits a property of an upper classification class; a proposal history storing unit that stores the past received proposal; an approximate proposal extracting unit that extracts the past received proposal stored by the proposal history storing unit approximate to the latest received proposal; and an approximate proposal presenting unit that presents the extracted proposal.

Disclosed in the invention is a client layered address list realization method for instant messaging. A directory server for providing an active directiory domain service, a UC server for providing a real-time collaborative service, and a database server are connected by networks to from an address list processing platform; and more than one client is connected with the database server and the UC server to carry out communication. A new user as one of contacts is established in a directory server; the new user is started in the UC server; and a whole contact-based enterprise organization architecture is established in the database server. The user only needs to install UC client software locally and logs in the software and can obtain the layered enterprise organization architecture information from the database server; classification can be carried out by a layered way and sub classification units can be established at all the classification units; and contacts can also be established, so that a layered communication list is established instead of a simple structure only with a classification-contact mode. Therefore, the simple and convenient using experience is provided for the user.

The present invention discloses a method and a system for ring tune uploading. At least one first level classification of the system ring tunes is established at the polyphonic ring gateway side, a first level classification of the system ring tunes which as much as the one at the polyphonic ring gateway side is pre-established at the polyphonic ring center side, at least first level sub-classification is established under the first level classification of the system ring tunes of the polyphonic ring center side, and when the system ring tunes needs to be upload, and the method comprises that the polyphonic ring gateway side upload the request according to the received system ring tunes, to get the upload information of the system ring tunes which is needed when uploading the system ring tunes, and send the upload information of the system ring tunes to the polyphonic ring center side; and the polyphonic ring center side puts the system ring tunes under the corresponding classification of the polyphonic ring center side according to the upload information of the system ring tunes. The ring tunes uploading failure caused by the shortage of the classified resource can be sharply reduced by pre-establishing a plurality of second level sub-classifications, and consequently the operation of the new adding classification at the polyphonic ring gateway side and the polyphonic ring center side is avoided.

The invention provides an electronic filing method and device for paper archives and terminal equipment, and the method comprises the steps: determining a first classification of a to-be-filed paper archive, and obtaining the image information of the to-be-filed paper archive; extracting keyword information of the to-be-archived paper archives based on the image information of the to-be-archived paper archives; detecting whether the title information of the to-be-archived paper archives is matched with the sub-classification names under the first classification or not, and detecting whether the keyword information of the to-be-archived paper archives is matched with keywords of electronic archives with the same sub-classification identification in an electronic archive basic information base or not; and archiving the to-be-archived paper archives according to the sub-classification name matching result and the keyword matching result. According to the electronic archiving method and device for the paper archives and the terminal equipment, the archiving efficiency of the paper archives can be improved.

The invention provides a text classification number determination method and device and electronic equipment, and relates to the technical field of computers. The method comprises the following steps:respectively inputting to-be-classified texts into a first classification model and a second classification model to obtain a first-level target classification number obtained through the first classification model and a second-level target classification number obtained through the second classification model; matching the first-level target classification number with the second-level target classification number; according to a matching result, determining at least one sub-classification model in a third classification model, the sub-classification model having a corresponding relationshipwith the target level classification number; respectively inputting the text to be classified into each sub-classification model, and determining a first parameter calculation result of each sub-classification model; and determining a target sub-classification model from the at least one sub-classification model according to the first parameter calculation result, and taking a target level classification number corresponding to the target sub-classification model as a classification number of the to-be-classified text. According to the invention, accurate classification of the to-be-classifiedtext can be efficiently realized.

The present invention provides a self-expanding method and system based on a cloud knowledge base. The method comprises: establishing a cloud knowledge base; creating a resource directory of the cloudknowledge base, wherein the resource directory is displayed through a classification list; according to the classifications in the classification list, establishing sub-classifications of each classification to form a classification tree, wherein the classification list comprises a plurality of multi-level classifications, at least an enterprise-level classification, a product-level classification, a project-level classification, and a code-level classification; and according to the classification tree, completing self-expansion of the cloud knowledge base. According to the technical scheme of the present invention, self-expansion and self-evolution can be carried out through accumulation of a large quantity of past project development, the knowledge that can be directly used by a programmer can be generated through a visual operation, and without repeated compilation, the development bottlenecks are broken through, so that the development efficiency is greatly improved, and the development workload of the programmer is reduced.

The invention discloses an accounting teaching demonstration system for a practical simulation module. The system comprises a simulation quantity import module, a random selection module, a sub-classification module, a standard quantity comparison module and a sorting output module, wherein the sub-classification module comprises a journal module, a trial table module, a day book module and a financial statement module that are distributed in parallel; and the sub-classification module transmits information to the standard quantity comparison module through a process A. According to the system, a conventional accounting teaching demonstration method is technically transformed through a modular classification and random sample selection combined method, so that the comprehensiveness and proficiency of skills of students can be greatly improved by using the system and the interests and learning effects of the students are greatly improved.

The embodiment of the invention discloses a video retrieval method and a video device, and relates to the technical field of electronic information. A video can be rapidly retrieved, and the video retrieval efficiency can be improved. According to the specific scheme, the video retrieval method includes the steps of obtaining a first video set; calculating information entropies of at least two attributive classifications in the first video set, wherein each attributive classification comprises at least two sub-classifications; reminding a user to carry out selection in the sub-classifications of the attributive classification with the largest information entropy. The video retrieval method and the video device are used in the video retrieval process.

The embodiment of the invention discloses a vulnerability scanning method and device based on log analysis, equipment and a storage medium. The method comprises the steps of obtaining a sample log file corresponding to sample vulnerability data, and analyzing the sample log file to obtain sample feature data; training a first sub-classifier according to the sample feature data set, wherein the first sub-classifier is used for outputting the probability that the sample log file corresponds to each vulnerability attack type; training a second sub-classifier according to an output result of the first sub-classifier, wherein the second sub-classifier is used for determining a vulnerability attack type corresponding to the sample log file according to the probability; integrating the first sub-classifier and the second sub-classifier into a vulnerability classifier; and inputting an analysis result of a to-be-detected log file into the vulnerability classifier for classification detection so as to confirm a vulnerability attack type corresponding to the to-be-detected log file. The vulnerability attack type can be quickly and accurately recognized, so that a targeted defense strategy isprovided.

The embodiment of the invention discloses a multi-view depth map enhancing system based on total probabilistic models. The system comprises the following steps: preprocessing an image; aggregating pixel vectors into superpixel vectors; classifying color information; utilizing a Dirichlet mixed model to simulate distribution of the superpixel vectors, and utilizing a variational method to estimate a model parameter to obtain a probability density function of the superpixel vectors, and carrying out classification according to the probability; classifying and enhancing depth information; utilizing a Gaussian mixture model to simulate distribution of the depth map pixel vectors, utilizing the variational method to estimate a model parameter, carrying out sub-classification on depth map pixels according to the obtained probability density function, updating depth values of the pixels into depth mean values in the categories of the pixels, and completing the depth map enhancement. By means of the multi-view depth map enhancing system based on total probabilistic models, a depth map cleaner in levels and compacter in structure can be obtained, so that the multi-view image synchronizing quality based on depth drawing is improved, and great practical value is achieved.

A method for concurrent diagnosis of multiple faults of an aeroengine. The total number of tags q is determined according to the number of sub-components of the engine object to be diagnosed, and thenthe total objective function is established according to the number. The objective function can be decomposed into q independent sub-problems, which can be solved by basic classification algorithm, such as SVM. When the engine training data set is class unbalanced, different penalty factors can be added to each sub-objective function based on SVM to keep the sub-classification boundary in the ideal position, A sub-classifier with good classification performance is established, which eliminates the defect of classic SVM that all errors are the same penalty coefficient, and it is inconvenient to adjust the sub-classifier boundary to the ideal position in the class imbalance data. The resulting q sub-classifiers are synthesized to obtain an overall classifier. The final overall classifier can diagnose both the concurrent instances of aero-engine multi-faults and the single fault instances of aero-engine.

The invention discloses a small sample multi-classification method based on a nested integrated depth support vector machine. According to the method, a two-layer integrated classification technologyis adopted to analyze multi-class classification tasks under the condition of small samples; for an internal layer, three different single-kernel functions (a radial basis kernel function, a Sigmod kernel function and a polynomial kernel function) are respectively given to a deep support vector machine (DSVM) of a sub-classifier, and a basis classifier is constructed through integrated classification; for an external layer, a'one-to-other 'strategy is used to convert a multi-class classification problem into a plurality of dichotomy problems. And finally, an AdaBoost framework is utilized to complete a multi-classification task. The method can effectively avoid an'over-fitting 'phenomenon under the condition of small samples, so that the accuracy of through-wall human body target recognition under the condition of small samples is improved.

The invention discloses a plasma rupture prediction method based on sub-classifier weight voting, and the method comprises the steps: 1, training N independent sub-classifiers based on a machine learning technology through N groups of independent plasma rupture feature data; 2, testing the prediction capability of each sub-classifier, and endowing each sub-classifier with a corresponding prediction weight; 3, distributing the feature data of the to-be-predicted time slice to each sub-classifier, and giving prediction result values by the sub-classifiers respectively; 4, carrying out weighted averaging on the prediction result values of the sub-classifiers to obtain a prediction coefficient R, and predicting the to-be-predicted time slice. According to the method, effective information of various fracture related signals can be effectively utilized, interference information is eliminated, the plasma fracture prediction accuracy is improved, and the false alarm phenomenon is reduced.

The invention provides a method for realizing multi-level classified management of digital photos on a digital TV and set top box equipment. The method comprises the following steps: the user sets different classifications for the photos in advance, each classification can be subdivided to set sub-classifications and the user stores the photos in different classification paths according to different photo contents. The technical scheme can realize multi-level digital photo classification and allow the user to carry out classification management on the digital photos, is convenient for the user to find the photos to be browsed, saves the operation time of the user and improves the efficiency.