44results about How to "Guaranteed search efficiency" patented technology

The invention discloses a method and equipment for compression and searching of index data. The method includes the steps that after an index information list is obtained, whether the column index data in every two adjacent lines and in the same column are the same is sequentially compared according to the sequence of column index data in the index information list; when the column index data in every two adjacent lines and in the same column are the same, the same column index data are compressed; when the column index data in every two adjacent lines and in the same column are different, the column index data with different comparison results in the every two adjacent lines and in the same column and corresponding position information are recorded, and the column index data in every two adjacent lines and in other columns after the same column where the column index data with different comparison results are located are compressed; after the column index data of each line in the index information list are compressed, an index array for the index information list is generated according to the compressed column index data, the memory capacity of the index array is reduced, the data amount is small, and the searching efficiency of the column index data is guaranteed.

The embodiment of the invention provides a document index generation method and device based on word segment weights. The method comprises following steps of setting anchor text weights corresponding to documents; configuring feature weights for word segments contained in the anchor texts according to the anchor text weights; determining the word segment weights of the word segments relative to the documents based on the feature weights; and generating document indexes according to the word segments and the word segment weights relative to the documents. According to the embodiment of the invention, the display probability of the search result terms related to user search is improved, the search accuracy is improved, further search carried out in modes such as turning pages to search in the search results and inputting the search keywords again is reduced, the operation simplicity is improved, the resource consumption of a search engine and a local system is reduced, the bandwidth consumption is reduced, and the search efficiency is improved.

The invention relates to a method for matching an IP address based on a suffix index. The method comprises the following steps: S1, constructing a suffix index for the IP routing data; S2, averagely dividing an IPv6 address into k segments, and then separately constructing different Hash tables for each segment of the IPv6 address, and ensuring that each Hash table stores interval values of different suffixes in the suffix index in the segment; S3, averagely dividing a given destination IP address into k segments, then searching the matched suffix index intervals for the k segments of the destination IP address in k Hash tables according to a binary searching method, and recording the number of segments corresponding to the Hash table where the matched suffix index interval is located; and S4, carrying out backtracking search in the segment according to the recorded number of segments, and outputting the routing information of the destination IP.

A method, a device and a system for a communication agent are used for performing switching between a physical interface and a logic interface on information of the physical interface and the logic interface in received messages according to obtained mapping information between the predetermined physical interface and the logic interface after the interactive messages between an operator device and an access device are received so as to isolate the physical interface between the operator device and the access device, accordingly changing of the physical interface can be ensured not to cause influence on the business development process of the operator device, and safety of various physical interfaces of the access device can be ensured.

The invention provides a method and a device for building log indexes. The method for building the log indexes includes determining at least one necessary field combination in log files according to preset field combination rules; executing operation for each log record in the log files, to be more specific, acquiring values of fields in the log records, splicing the values of the fields to obtainvalues of the necessary field combinations and acquiring offset locations of the log records in the log files; creating index records for the different values of each field combination; writing the index records of the same field combination into a single index file of the corresponding log file. The index files need to be created for the necessary field combinations determined according to the preset field combination rules. The index records further comprise offset locations of the log records corresponding to the different values. According to the scheme, the method and the device have theadvantage that the log searching efficiency can be effectively improved by the method and the device.

The invention relates to a method for matching IP addresses based on a suffix index, comprising the following steps: S1. constructing a suffix index for IP routing data; S2. dividing the IPv6 address into k segments on average, and then constructing different hashes for each segment of the IPv6 address Hash table, each hash table stores the interval values ​​of different suffixes in the suffix index in the segment; S3. For a given destination IP address, divide it into k segments on average, and then use the binary search method for the destination IP address of k segments Find the matching suffix index interval in the k hash tables, and record the segment number corresponding to the hash table where the matching suffix index interval is located; S4. Perform an intra-segment backtracking search according to the recorded segment number, and output the routing information of the destination IP.

The invention relates to the technical field of image processing, in particular to an image searching method and a device, a a storage medium and electronic equipment. According to the image searching method provided by the embodiment of the invention, an image search request sent by a client based on a panoramic compressed image can be received, the image search request carries selection box information, and a corresponding target object compressed image is intercepted from the panoramic compressed image based on the selection box information, compressed image features of the target object compressed image are extracted, the matching degrees between the compressed image features and original image features of all pre-stored local original images is obtained, and target original image features of which the matching degrees meet a preset matching condition is selected based on all the obtained matching degrees; and image search is carried out based on the target original image feature to obtain a corresponding search result, so that the accuracy of the image search result can be improved.

The invention discloses a two-stage particle swarm optimization algorithm including independent global search; it includes the steps of initializing the population; adopting a chaotic method to initialize the position X and velocity V of the particles; using the fitness function fitness to calculate the current values ​​of all particles Adaptation value, initialize the historical optimal position pbesti of each particle and the global optimal position gbest of all particles; perform the first-stage iteration-global search; perform the second-stage iteration-local search. Beneficial effects of the present invention: in each iteration of the first stage iteration-global search, the particle randomly selects a non-self particle to learn from all the particles, and this random selection ensures that the group will not track specific particles, and also This ensures that aggregation does not occur. The second stage of iterative-local search can quickly converge and obtain a high-precision solution, the precision of the optimal solution is improved, and the "premature" defect is significantly improved.

A method for planning and dynamic updating of an underwater target search path coordinated by an ROV and a mother ship, comprising the following steps: calculating a path point of the surface mother ship according to a provided search area, and planning the motion path of the surface mother ship; the surface mother ship moves to the way point; Under the condition that the relative relationship between the ROV and the surface mother ship is normal, the current search sub-area is rasterized for the first time. The first layer of grid is the ROV global path planning layer. The biological excitation neural network algorithm is used for the first layer of grid. Carry out global path planning and dynamic update; the first layer of grid is rasterized for the second time, and the second layer of grid is the ROV path dynamic update layer, which realizes the local trajectory planning and dynamic update of the ROV, and judges whether the ROV is triggered in real time. The second-layer grid dynamic update condition is set until the current first-layer grid search is completed. This design achieves full coverage of the search area by forward-looking sonar while ensuring the safe operation and search efficiency of ROV.

A configurable intelligent optimization method based on a double-layer automatic generation strategy combines an operator configuration method in a serial intelligent optimization method and a topological structure configuration mode in a parallel intelligent optimization method, and divides sub-populations in an evolutionary iteration process through the double-layer automatic generation strategy. Self-adaptive communication topology selection is realized through automatic operator configuration and inter-group interaction, synchronous search of a plurality of processes or threads is formed, and diffusion and application of population overall information are realized in a periodic process interaction mode, so that the global optimization efficiency is ensured under the condition of sub-population dispersion independent solution, the dynamic nature and robustness of the intelligent optimization method in large-scale solution space search are enhanced, complex problem decomposition and optimization are more flexible, and more space is provided for hybrid design and solution of the intelligent optimization method.

The invention relates to a method of improving data storage efficiency and security of an intelligent network load interaction terminal, and belongs to the technical field of network communication. The method includes the steps of a storage step, an extraction step, regular reading, data recovery and network asking for help. Logical files are stored in physical files in a manner of gathering partsinto a whole, and thus storage capacity is significantly increased; and a first-level index file and a second-level index file are also introduced, and thus finding efficiency of data is effectivelyguaranteed. In addition, a multi-guarantee mechanism of CRC32 checking, compressed backing-up and network asking for help is also adopted, the method has extremely high data recovery capability, and the security of data storage and reliability of terminal running are ensured.

The invention discloses a method and a device for searching a target image. The method comprises the following steps: collecting a plurality of feature points in a target subimage; converting the target image into a one-dimensional array from a two-dimensional pixel matrix, and adjusting relative offsets among the feature points corresponding to the conversion manner; searching the one-dimensional array by using the adjusted relative offsets. On one hand, the feature points are adopted for searching, thus the defects of a method for contrasting pixels frame by frame in a correlation technique can be avoided, and on the other hand, the one-dimensional array corresponding to the feature points is adopted, thus the calculated quantity in the searching can be reduced further, and the search efficiency is ensured.

Provided is an AI model automatic generation method based on computational graph evolution. The method mainly comprises the following steps: pre-setting data; utilizing a genetic algorithm operator to generate a first-generation computational graph model, and computing the performance of the model according to a computational graph structure thereof; removing an invalid model and a repeated model, and taking the remaining models as candidate models and reserving same as seeds for the next generation; selecting a number of optimal models; the candidate models generating a new computational graph model by using the genetic algorithm operator; determining whether the new computational graph model generated in the last step has been generated; storing the new model as a computational graph model for a new generation, and determining whether same satisfies the pre-set data and an evolution ending condition; and summarizing evolution computational results, and selecting an optimal model. In the present invention, machine learning and deep learning can be carried out simultaneously; the repeated computation of the same model is prevented, and the model design efficiency is improved; the local optimum is jumped out of; the decline in the ability to search for a network is prevented; and evaluation can be directly carried out without training by means of actual data.

The invention discloses a method for realizing a real-time database for a miniature power grid intelligent terminal. The method comprises the following steps of: mapping data sheets, regions and records forming a real-time database to internal storages of the miniature power grid intelligent terminal in a file mapping mode, and forming a database entity; and storing the head address of a tablespace and the address of a record pointer in a corresponding B+ tree. The real-time database generated according to the method can operate in the miniature power grid intelligent terminal which is only provided with an embedded processor and few internal storages; the data access of the real-time database can be kept in a microsecond speed; and complex and dynamic variation data can be stored in the real-time database.

A method, device and system for communication proxy are provided. After receiving an interactive message between a provider device and an access device, performing a converting process between a physical port and a logical port on the physical port information or logical port information in the received message, according to the obtained predetermined mapping information between the physical port and the logical port, thereby the physical port between the provider device and the access device is isolated.

The invention discloses a method for predicting an octane loss value based on a random forest parameter optimized by particle swarm optimization. 2. Preprocess the remaining data; step 3, use the training data set to train the random forest algorithm to obtain a training model; step 4, initialize the parameters of the particle swarm algorithm; step 5, take the root mean square error as the adaptation of the particle swarm algorithm degree function, the number of parameter decision trees in the training model and the depth of the tree are continuously solved by the particle swarm algorithm, and the optimal parameters are imported into the training model to obtain the optimal prediction model; Step 6, re-input the new test set to import The optimal prediction model is tested and the prediction result is obtained. The present invention can be effectively used for octane loss prediction.

The present invention relates to a font similarity and font replacement method, comprising: step 1, respectively constructing Chinese character text block images according to preset attribute values ​​that affect the visual effect of fonts, so as to generate a Chinese character text block image sample library; wherein the Chinese character The text block image includes n×n non-overlapping Chinese characters, where n≥2; Step 2, calculate the texture feature value of the Chinese character text block image set; Step 3, use the static replacement method or dynamic replacement method to replace the missing font .