30 results about "Numerical types" patented technology

Data mining techniques are provided which are effective and efficient for discovering useful information from an amorphous collection or data set of records. For example, the present invention provides for the mining of data, e.g., of several or many records, to discover interesting associations between entries of qualitative text, and covariances between data of quantitative numerical types, in records. Although not limited thereto, the invention has particular application and advantage when the data is of a type such as clinical, pharmacogenomic, forensic, police and financial records, which are characterized by many varied entries, since the problem is then said to be one of “high dimensionality” which has posed mathematical and technical difficulties for researchers. This is especially true when considering strong negative associations and negative covariance, i.e., between items of data which may so rarely come together that their concurrence is never seen in any record, yet the fact that this is not expected is of potential great interest.

The invention discloses a data storage and search method based on numerical indexing of spatial data. The data storage and search method comprises the steps as follows: 1) automatically acquiring a page with the spatial data from the internet, analyzing and storing in a database; 2) selecting properties to be indexed from the database, converting the non-numerical type properties in the database into corresponding numerical values, and storing; 3) carrying out classified indexing on the spatial data according to spatial data types, creating a numerical indexing file for a data table of each data product, taking the numerical indexing file as a main index, creating an increment index according to the set time cycle, and periodically combining the increment index with the main index, wherein the numerical indexing file comprises an indexed property and inverted file mapping table and a plurality of numerical document inverted lists; 4) deploying the indexes to a search service end; and 5) searching the data table by the search service end according to a received search request. According to the data storage and search method, the search speed is increased, and incremental indexing is supported.

The invention discloses a multidimensional information fusion-based comprehensive e-commerce product scoring method. The method comprises the following steps of: obtaining multidimensional informationsuch as shop information, sales volume information and comment text information of e-commerce products; data preprocessing: carrying out data cleaning and data conversion on numerical type data, andcarrying out word segmentation and part-of-speech tagging on comment texts; mining the multidimensional information: carrying out data reduction and principal component regression analysis on the shopinformation and the commodity sales volume information to obtain shop information indexes and commodity sales volume indexes, carrying out emotion analysis on the comment texts, and obtaining a product feature score radar map through a quantification method and a clustering method; and commodity total score calculation: designing a fusion function and calculating a commodity total score. The method can be applied to commodity information-based commodity recommendation systems, is capable of efficiently and conveniently recognizing high-quality commodities so as to ensure that the designed recommendation systems are more rapid and correct.

The invention discloses a circuit and a method for realizing data sorting and solves the problems of current software that the sorting time is long and the real-time request is not met. The circuit and the method of the invention are that: determining the depths of a register group, a comparator group, a multiplexer group of one chosen from n, a first multiplexer of one chosen from two, a second multiplexer and an extreme-valued pointer register and the widths of the first multiplexer of one chosen from two, the second multiplexer and the extreme-valued pointer register according to the extreme value number n which is needed to be searched; determining the running time of the circuit as m hours according to the number m of a data source; choosing an output type of the comparator according to a searching numerical type; resetting the register group and the extreme-valued pointer register group; inputting a sampled data to the circuit per clock cycle; stopping the circuit after m hours when the extreme value is n which is preserved in the register group. The circuit and the method of the invention are characterized in that: the real-time processing of the circuit is strong and the sorting time is reduced by times.

The present invention relates to an auto associating and analyzing cloud computing monitor apparatus and method. The system comprises a configuration push component, a data collation module, and an associating and analyzing system. The configuration push component is used for monitoring proxy configuration templates to be allocated to related monitored devices in a unified manner. The data collation module is used for collecting monitoring data at the same frequency regularly so as to generate monitoring items and host status data, and collating character string data so as to generate numerical type or character type monitoring values. The associating and analyzing system is used for associating and analyzing the monitoring data by time period, and providing associated data of a timeline according to a trend curve of a time associated displayed monitoring items, and downtime and alarming quantity of the same time period, so as to provide auxiliary information for fault analysis. Monitoring configuration is delivered in a centralized manner, monitoring data is collected automatically, comprehensive system running indication and alarm prompt is provided for a maintenance person, and the fault handling efficiency is improved through auto associating and analyzing.

The invention relates to an incomplete data fuzzy clustering method for information feedback RBF network estimations, which comprises the following steps: 1) presenting an information feedback RBF network model; 2) presenting an incomplete data fuzzy clustering method (IFRBF-FCM) of information feedback RBF value estimations; 3) selecting the corresponding training sample set for the incomplete data sample by using the nearest neighbor rule, and training the IFRBF network for each missing attribute by using the nearest neighbor training sample set, thereby realizing the estimation prediction of the missing attribute in the incomplete data sample and obtaining the complete data set after the estimation recovery of the IFRBF network; 4) determining the estimation interval of the attribute ofthe incomplete data to propose an incomplete data fuzzy clustering method (IFRBF-IFCM) of IFRBF interval estimations to obtain fuzzy clustering results. The invention adopts the IFRBF network to estimate the incomplete data set and recovers the intact data set. Compared with the comparison method, the clustering result of the intact data set is more accurate than that of numerical type estimations, and the robustness is better.

The invention discloses a method for semi-supervised learning of structured data, and the method comprises the steps: building a semi-supervised adversarial neural network model for the structured data, carrying out the preprocessing of original structured data X, and enabling the features of the original data X to be divided into a category type feature subset xCT and a numerical type feature subset xNL; the original input of the model discriminator is {x1, x2, xg}, wherein x1 is a positive integer;, xu is respectively a labeled sample and an unlabeled sample; wherein xg is a sample generatedby the generator; feature sets contained in x1 and xu are the same; inputting the class feature subset xCT of the sample into an Engineering layer; obtaining a corresponding dense embedding vector E(xCT), combining the dense embedding vector E (xCT) with the numeric feature subset xNL to obtain a sample E(xCT) + xNL with a new feature set, obtaining a normalized sample containing the new featureset by applying a BN technology, inputting the new sample into a discriminator for training, and generating a sample xg which is directly used as the input of the discriminator; the generator is composed of three layers of full-connection networks, BN is applied to the output of each layer to prevent gradient dispersion, noise serves as the output, and a production sample xg with the characteristic E(xCT) + xNL is obtained.

The invention discloses a search method and device. The search method includes that receiving search terms; analyzing the search terms and extracting numerical type demand information from the search terms; acquiring a search result according to the search terms and extracting an abstract from the search result according to the numerical type demand information to generate the abstract corresponding to the search result; providing the search result and the abstract corresponding to the search result. The search method is capable of meeting the search requirements of a user on semantics and also capable of ensuring the relevance to the search terms from the text.

The invention provides an abstract memory model-based method for calculating non-numerical type data. The method comprises the following steps of: A, designing an abstract memory model, wherein the abstract memory model is used for simulating a memory structure of numerical type variants and non-numerical type variants, and storing semantic information and restraint relationship included in the variant operation; B, extracting the semantic information included in the type operation of the numerical type variants and the non-numerical type variants, and mapping the semantic information to an abstract memory model; C, extracting restraint among variants and restraint inside variants included in the type operation of the numerical type variants and the non-numerical type variants, and mapping the restraint relationship to the abstract memory model; and D, extracting the semantic information and the restraint relationship of the variants from the abstract memory model, and establishing a test case by using a test case establishing algorithm and a restraint solver of a third party. By utilizing the method, the defect that in the prior art the non-numerical type variant program semantic cannot be supported is overcome, and the purpose that a program including the non-numerical type automatically generates the test case is realized.

The embodiment of the invention provides a data processing method and device, and the method can comprise the following steps: carrying out the standardized coding of input table data, obtaining firsttable data, and an object description feature of the first table data being a numerical type object description feature; based on the first table data, using a generative adversarial network model togenerate second table data, and the similarity between the second table data and the first table data reaches a first threshold value; and performing inverse standardized encoding on the second tabledata to obtain output table data, the output table data having the same object description feature as the input table data. By adopting the embodiment of the invention, the output data very close tothe input data can be constructed, and the data analysis can be realized even if the data is separated from the data local point.

The present invention provides a method and apparatus for ciphertext retrieval of numerical data based on cloud storage. The method includes: generating a key sequence according to the key of an OAIS data packet to be retrieved, and splitting the key sequence into a plurality of value sequence values with set lengths; translating the plaintext query sentence of the OAIS data packet to be retrieved into a corresponding ciphertext query sentence, wherein a plaintext searching condition that the user inputs is included in the plaintext query sentence, and a ciphertext searching condition obtained by the plaintext searching condition is included in the ciphertext query sentence; submitting the equivalent searching condition corresponding to the ciphertext searching condition to the cloud server to retrieve and then obtaining the retrieval result; mapping the identification of the retrieval result in the database to obtain the corresponding identification of the sequence values; extracting the corresponding identifications of all the sequence values which are larger or smaller than the mapped sequence value to obtain a identification set; obtaining the corresponding data from the cloud server according to the identification set. The invention can improve the ciphertext retrieval efficiency of the cloud storage on the basis of ensuring data security.

The invention discloses a deserialization method and device, and relates to the technical field of data processing, which can solve the problem that in the process of deserializing a JSON string intoa Java object, when the data content of the original data type being numberical type in the JSON string is deserialized according to the numeric type of the data content with the highest data precision, the data precision of some data content in the obtained deserialized Java object changes in the prior art. The method of the invention comprises the following steps of: receiving a JSON string; parsing the JSON string to obtain the parsing result; determining an original data type of a data content corresponding to each field name, and judging whether the data content contains at least two datacontents of which the original data type is a numerical type; If so, the JSON string is converted to a Java object according to the original data type corresponding to each data content. The invention is suitable for application in the process of deserializing a JSON string into a Java object.

The invention discloses an automatic feature construction method based on an attention mechanism and reinforcement learning, and the method sequentially comprises the following steps: 1, giving a dataset DTR of a classification problem, including a numerical type feature set S, setting the parameter maximum iteration number maxIterations, and setting the value of an embedded size as embeddingSize; and 2, transmitting the data set and the parameters into the automatic feature construction method, and operating to obtain a classification result. According to the method, a feature generator based on a self-attention mechanism and a feature selector based on reinforcement learning are included, generated features are continuously explored and utilized through iteration, feature generation ofa test set is guided through a globally optimal feature generation and selection scheme in limited steps, and therefore an optimal classification result is automatically obtained.

The invention discloses a document parsing method, system and device. The method includes: utilizing regular expression rules of financial indexes, starting characteristic indexes and ending characteristic indexes to carry out locating of financial statements on different-format documents; when data in the financial statements are located, recording the financial data and corresponding index namesand times; and carrying out unit conversion on numerical-type data, and then recording the data. The system includes a construction unit, an acquisition unit, a first locating unit, a second locatingunit and a conversion unit. The device includes a memory and a processor. When a program is executed by the processor, the processor is enabled to realize the document parsing method. According to the method, system and device, the financial data in the different-format documents can be quickly and accurately parsed, and application compatibility, comprehensiveness, accuracy, and processing efficiency of a parsing solution are improved. The document parsing method, system and device of the invention can be widely applied in the field of big-data parsing technology.

The invention provides a system-level performance data processing method and equipment utilizing the same. The system-level performance data processing method includes: acquiring performance data corresponding to assigned performance index marks from each network element of a network element set; judging whether the type of the acquired performance data is percentage type or numerical type, averaging the acquired performance data if the type of the acquired performance data is the percentage type, utilizing the obtained average value as integral performance data and adding the integral performance data to a performance table; summing the acquired performance data if the type of the acquired performance data is the numerical type, utilizing the obtained sum value as the integral performance data, and adding the integral performance data into the performance table. System-level performance data can be acquired by the system-level performance data processing method the equipment utilizing the same.

The invention discloses a method for detecting abnormal behaviors of the Internet of Things based on a machine learning technology. The method comprises the following steps: A, collecting protocol data of the Internet of Things; b, performing feature processing on the collected data, and distinguishing whether the data belongs to a numerical type or a character type according to key values in thedata; if the character type belongs to the numeric type, entering the step C, and if the character type belongs to the character type, entering the step D; C, performing periodic behavior calculationon the data belonging to the numerical type, modeling by utilizing a machine learning algorithm, and predicting abnormal data; and D, performing word segmentation processing on the data belonging to the character type to judge the character entropy, performing machine learning algorithm modeling, and calculating outliers so as to distinguish abnormal data. According to the method, the normal flowcharacteristics of the Internet of Things are analyzed by utilizing a machine learning algorithm, so that the flow behavior is predicted.

The invention discloses a numerical type financial data quality monitoring system and method. The system comprises a data reading configuration module to be monitored, a data reading and standardization module, a monitoring model configuration module, a data monitoring rule self-generation and update module, a data monitoring rule manual configuration module, a scheduling module, a data monitoringcalculation module, a message pushing module and a feedback module. The numerical type financial data quality monitoring system and method provided by the invention adopts machine learning to carry out self-learning on monitoring rules, and combines the manually set data quality monitoring rules with the machine self-learning monitoring rules, so as to improve the quality, efficiency and coverageof data quality monitoring, and reduce the human cost at the same time.

The invention provides a method for storing a plurality of numerical value rows into one line, which defines field data structures of the stored numerical value rows as a non-numerical type, i.e. varchar or text and then connects ideographic strings of numerical values by '-' and ',' through program analysis and stores the ideographic strings into a database.

The invention discloses a Chinese-wolfberry data identification method. The method includes: step 1, obtaining Chinese-wolfberry information data, wherein the Chinese-wolfberry information data include numerical-type data and non-numerical-type data; step 2, retrieving archives, which contain the Chinese-wolfberry information data, according to the numerical-type data and the non-numerical-type data, selecting archives matching the numerical-type data and the non-numerical-type data in the Chinese-wolfberry information data, and filling a database with matching results; and step 3, if not, using the Chinese-wolfberry information data to supplement the database to form a new archive. The method has the advantages of: improving speed of identification, correcting errors contained by Chinese-wolfberry information in the database, such as modifying wrongly written characters in the database, and improving the speed and accuracy of identification.

The invention relates to a digital watermark copyright protection method for numerical relational databases, belongs to the field of computer security, and specifically provides a method for embedding, extracting and verifying watermarks of numerical relational databases. The method includes pattern and data analysis of numerical relational database, generation and embedding of digital watermark, and detection method of digital watermark. Aiming at the characteristics of numerical relational databases, the invention can efficiently and dynamically inject digital watermark information into data evenly, and extract digital watermark information without source data. The method adopts technologies such as key encryption and anti-reordering, which can effectively resist various types of watermark removal attacks, and can ensure that the watermark will not be lost when less than 50% of the rows are modified.

The invention relates to a DBSCAN clustering algorithm-based method for checking missing registration of important personnel and houses, which comprises the following steps of: preprocessing population and house data sets collected by civil police, including missing value filling, category type variable discretization and numerical type variable standardization; adopting a DBSCAN clustering algorithm to classify samples of non-core points on a 'key persons and houses' data set, and analyzing the clustering result; fixing core points of data of all labels' key persons and houses', classifying samples of non-core points on a population and house data set through a DBSCAN clustering algorithm with adaptive characteristic weights, obtaining a clustering result, and finally generating a suspected missed registration 'important person and house' check table. Therefore, the labeled key concerned persons and houses are taken as the core, persons and houses similar to the key persons and the houses are screened out through the density clustering algorithm, then the checking range of suspected key persons and houses is narrowed, and the police checking precision and efficiency can be effectively improved.

The invention discloses a repeated push-change type calendar which comprises two parts, namely, a main panel and a push-pull plate. The main panel is provided with a numerical type data scale, a square display mouth is formed in the middle portion of the push-pull plate, a week scale is arranged above the square display mouth, a month indicating arrow is arranged on the right side of the square display mouth, and the week scale is matched with the data scale. A clamping groove is formed in the main panel, and the push-pull plate is installed in the clamping groove of the main panel. The repeated push-change type calendar uses the law of the arrangement of a calendar, the date and week combination of a month is arranged on the square display mouth, and date and week are enabled to correspond through appropriate pushing of the square display mouth of the push-pull plate. The display of all months can be achieved, the changing into a next month can be achieved by pushing the push-pull plate, and therefore the repeated using of the calendar is achieved, and energy-saving and environment-friendly effects are achieved.

The invention discloses a data storage and search method based on numerical indexing of spatial data. The data storage and search method comprises the steps as follows: 1) automatically acquiring a page with the spatial data from the internet, analyzing and storing in a database; 2) selecting properties to be indexed from the database, converting the non-numerical type properties in the database into corresponding numerical values, and storing; 3) carrying out classified indexing on the spatial data according to spatial data types, creating a numerical indexing file for a data table of each data product, taking the numerical indexing file as a main index, creating an increment index according to the set time cycle, and periodically combining the increment index with the main index, wherein the numerical indexing file comprises an indexed property and inverted file mapping table and a plurality of numerical document inverted lists; 4) deploying the indexes to a search service end; and 5) searching the data table by the search service end according to a received search request. According to the data storage and search method, the search speed is increased, and incremental indexing is supported.

The invention discloses an implementation method of a general input edit control suitable for an intelligent instrument, which belongs to the field of intelligent instruments. The implementation method comprises the following steps of (1) defining a value type, physical significance and continuity or discretization; (2) defining a keypress response algorithm; (3) defining attributes of input parameters; (4) defining the format of an instrument parameter response function. According to the implementation method of the general input edit control suitable for the intelligent instrument, a consistent optional attribute definition input mode is designed, the input attributes are defined for combination, a numerical index is input, and input response functions and function lookup modes are normalized, so that various types of integer numerical value input, floating-point numerical value input, symbol input and input association are integrally realized, software development and maintenance are facilitated, the development and maintenance efficiency of the instrument is improved, the operating efficiency is improved, and the cost is low.

The application discloses a method for evaluating the state of a transformer. The method includes the following steps that: continuous numerical type five-dimensional feature attributes are combined with one-dimensional type class labels to establish a historical knowledge base, wherein the five-dimensional feature attributes include a load rate, three-phase unbalance degree data, a winding temperature rise, an insulation oil temperature rise, and vibration data and the one-dimensional type class labels include a transformer state; a split point is set and the continuous numerical type five-dimensional feature attributes are discretized into Boolean attributes to complete data preprocessing; a decision-making classification model based on Gini index attribute selection measurement is established; on the basis of a cross training method and a self-help training method, the model is evaluated and optimized to obtain a new decision-making classification model; and the state of a transformer is evaluated based on the new decision-making classification model. Comprehensive evaluation is carried out by using the five-dimensional attributes of the electrical quantities including the loadrate and the three-phase unbalance and the non-electrical quantities including the winding temperature rise, insulating oil temperature rise and vibration frequency; and the evaluation parameters become comprehensive based on the decision-making classification method of the Gini index attribute selection measurement, so that the evaluation accuracy is improved effectively.

The invention discloses a numerical type sensor data filtering and transmitting method oriented to a crowd sensing application. The method comprises the following steps: 1) acquiring data through a sensor, and storing the data in a cache; 2) processing the data newly added in the cache one by one by use of a moving average method so as to filter random error of the data; 3) judging the size of the current cache, and executing the following step 4) if the size of the current cache achieves a set threshold value, or returning to the above step 1); 4) writing the data in the cache in a file to store, and then emptying the cache; 5) computing the time interval between the former transmission and the current, and executing the step 6) if the time interval achieves the set threshold value, or returning to the step 1); and 6) naming the file in the step 4) by use of the current time, and transmitting the file to a server-side according to a set grouping transmission strategy. Based on the crowd sensing environment, the method can be generally applied to the crowd sensing application of the numerical type sensor, and the practicability and the universality are realized.