CN111414451A - Information identification method and device, computer equipment and storage medium - Google Patents
Information identification method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111414451A CN111414451A CN202010124754.3A CN202010124754A CN111414451A CN 111414451 A CN111414451 A CN 111414451A CN 202010124754 A CN202010124754 A CN 202010124754A CN 111414451 A CN111414451 A CN 111414451A
- Authority
- CN
- China
- Prior art keywords
- user
- text information
- underwriting
- area
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3343—Query execution using phonetics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0281—Customer communication at a business location, e.g. providing product or service information, consulting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- General Business, Economics & Management (AREA)
- Economics (AREA)
- Marketing (AREA)
- Human Computer Interaction (AREA)
- Technology Law (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Acoustics & Sound (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The application relates to the technical field of classification models, and provides an information identification method, an information identification device, computer equipment and a storage medium, wherein the information identification method is used for receiving voice information returned by a target user and carrying out voice conversion on the voice information to obtain corresponding user text information; respectively inputting user text information into a preset recognition model; the recognition model is obtained by fusing a CNN model and a DBM model; extracting local characteristic vectors of the user text information by the full connection layer of the CNN model, and extracting global characteristic vectors of the user text information by the DBM model; fusing the local feature vectors and the global feature vectors to obtain fused feature vectors and inputting the fused feature vectors to a classification layer; and finally, carrying out classification calculation on the fusion feature vectors through a classification layer to obtain a classification result corresponding to the user text information. According to the method and the device, local features in the user text information are considered, and global features are combined, so that the final classification result is more accurate.
Description
Technical Field
The present application relates to the field of classification model technology, and in particular, to an information identification method, apparatus, computer device, and storage medium.
Background
At present, based on the increasing of the human cost of the agent, each company needs to reduce the human cost urgently, an intelligent robot is adopted to perform information interaction of each product, whether a user has a corresponding requirement or not needs to be identified in the process of performing information interaction with the user, the user behavior can still be identified only manually at present, the intelligent robot cannot be used for identifying whether the user has a corresponding requirement or not from answer information of the user, the efficiency is low, and the human cost is high; meanwhile, during manual identification, different agents have different subjectivity, so that the identification results are different, and the real requirements of the user cannot be accurately identified.
Disclosure of Invention
The application mainly aims to provide an information identification method, an information identification device, computer equipment and a storage medium, and aims to overcome the defect that user requirements cannot be accurately identified from answer information of a user at present.
In order to achieve the above object, the present application provides an information identification method, including the steps of:
reading the continuous underwriting years of each user stored in a preset database, and selecting a target user with the continuous underwriting years larger than a preset value;
outputting preset sales text information to a user terminal where the target user is located;
receiving voice information returned by the target user through the user terminal, and carrying out voice conversion on the voice information to obtain corresponding user text information;
inputting the user text information into a preset recognition model respectively; the recognition model is obtained by fusing a CNN model and a DBM model, wherein a classification layer which is formed by connecting a full connection layer of the CNN model and a characteristic output layer of the DBM model together is used as a final output layer of the recognition model;
extracting local feature vectors of the user text information by a full connection layer of the CNN model, and extracting global feature vectors of the user text information through the DBM model;
fusing the local feature vector and the global feature vector to obtain a fused feature vector and inputting the fused feature vector to the classification layer;
and carrying out classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, wherein the classification result is used as a recognition result of the voice information returned by the target user.
Further, after the step of performing classification calculation on the fusion feature vector through the classification layer to obtain a classification result corresponding to the user text information as a recognition result of the voice information returned by the target user, the method includes:
creating a blank layer;
acquiring the sales text information and the number of characters of the user text information;
calculating the size of a first area and the size of a second area correspondingly occupied by the sales text information and the user text information added into the blank layer according to the sales text information, the number of characters of the user text information and a preset font format;
dividing a first area corresponding to the size of the first area and a second area corresponding to the size of the second area in the blank layer; dividing a third area with a preset size between the first area and the second area in the blank layer;
adding the sales text information into the first area, adding the user text information into the second area, adding the classification result into the third area, and synthesizing into a picture.
Further, the step of adding the sales text information to the first area, the user text information to the second area, and the classification result to the third area includes:
performing hash calculation on the sales text information, the user text information and the classification result respectively to obtain a corresponding sales hash value, a corresponding user hash value and a corresponding classification hash value;
encrypting the sales text information through the classified hash value, and adding the encrypted sales text information into the first area;
encrypting the user text information through the sales hash value, and adding the encrypted user text information into the second area;
and encrypting the classification result through the user hash value, and adding the encrypted classification result into the third area.
Further, the classification result comprises an underwriting label and a corresponding prediction probability; after the step of performing classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, and taking the classification result as a recognition result of the voice information returned by the target user, the method includes:
saving the sales text information and the user text information in the preset database;
combining the sales text information with the underwriting label to obtain a first combination;
combining the user text information with the prediction probability corresponding to the underwriting label to obtain a second combination;
respectively calculating the first combination and the second combination through a Hash algorithm to obtain a corresponding first Hash value and a corresponding second Hash value;
adding the first hash value and the second hash value into the same array to generate a first collection; and uploading the first collection to an enterprise underwriting management server for storage.
Further, before the step of reading the continuous underwriting years of each user stored in the preset database and selecting the target user with the continuous underwriting years greater than the preset value, the method includes:
acquiring historical insurance policy underwriting data of a user based on a spark data calculation engine;
calculating the historical insurance policy underwriting data of the user to obtain the underwriting years of the user;
sequencing the underwriting years based on an array sort algorithm to obtain an underwriting year sequence which is arranged from big to small;
sequentially calculating differences between adjacent underwriting years in the underwriting year sequence, and sequentially forming the differences into a difference sequence;
correspondingly inputting the numerical values in the difference value sequence into a preset formula to calculate the continuous underwriting years of the user;
and storing the continuous underwriting years of the user and the user information of the user into a preset database in a correlation manner.
Further, after the step of correspondingly inputting the numerical values in the difference value sequence into a preset formula to calculate the continuous underwriting years of the user, the method includes:
acquiring the years of the user underwriting according to the underwriting years of the user;
acquiring the year span of the underwriting of the user according to the maximum year and the minimum year in the underwriting years of the user;
calculating the continuous underwriting probability of the user according to the years of underwriting of the user, the continuous underwriting years and the year span;
the classification result comprises an underwriting label and a corresponding prediction probability, and after the step of performing classification calculation on the fusion feature vector through the classification layer to obtain a classification result corresponding to the user text information as a recognition result of the voice information returned by the target user, the method comprises the following steps of:
and carrying out weighted calculation according to the continuous underwriting probability and the prediction probability corresponding to the underwriting label to obtain the underwriting probability of the user.
The present application further provides an information recognition apparatus, including:
the selection unit is used for reading the continuous underwriting years of each user stored in the preset database and selecting a target user with the continuous underwriting years larger than a preset value from the continuous underwriting years;
the output unit is used for outputting preset sales text information to a user terminal where the target user is located;
the conversion unit is used for receiving voice information returned by the target user through the user terminal and carrying out voice conversion on the voice information to obtain corresponding user text information;
the input unit is used for respectively inputting the user text information into a preset recognition model; the recognition model is obtained by fusing a CNN model and a DBM model, wherein a classification layer which is formed by connecting a full connection layer of the CNN model and a characteristic output layer of the DBM model together is used as a final output layer of the recognition model;
the extraction unit is used for extracting the local feature vector of the user text information from the full connection layer of the CNN model and extracting the global feature vector of the user text information through the DBM model;
the fusion unit is used for fusing the local feature vector and the global feature vector to obtain a fusion feature vector and inputting the fusion feature vector to the classification layer;
and the classification unit is used for performing classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, and the classification result is used as a recognition result of the voice information returned by the target user.
Further, still include:
the creating unit is used for creating a blank layer;
a first acquiring unit, configured to acquire the sales text information and the number of characters of the user text information;
the first calculation unit is used for calculating the size of a first area and the size of a second area correspondingly occupied by the sales text information and the user text information added into the blank layer according to the sales text information, the number of characters of the user text information and a preset font format;
the dividing unit is used for dividing a first area corresponding to the size of the first area and a second area corresponding to the size of the second area in the blank layer; dividing a third area with a preset size between the first area and the second area in the blank layer;
and the adding unit is used for adding the sales text information into the first area, adding the user text information into the second area, adding the classification result into the third area and synthesizing the classification result into a picture.
Further, the adding unit includes:
the calculation subunit is configured to perform hash calculation on the sales text information, the user text information, and the classification result, respectively, to obtain a corresponding sales hash value, a corresponding user hash value, and a corresponding classification hash value;
the first adding subunit is used for encrypting the sales text information through the classified hash value and adding the encrypted sales text information into the first area;
the second adding subunit is used for encrypting the user text information through the sales hash value and adding the encrypted user text information into the second area;
and the third adding subunit is configured to encrypt the classification result according to the user hash value, and add the encrypted classification result to the third area.
Further, the classification result comprises an underwriting label and a corresponding prediction probability; the information recognition apparatus further includes:
the storage unit is used for storing the sales text information and the user text information in the preset database;
the first combination unit is used for combining the sales text information and the underwriting label to obtain a first combination;
the second combination unit is used for combining the user text information with the prediction probability corresponding to the underwriting label to obtain a second combination;
the second calculation unit is used for calculating the first combination and the second combination respectively through a Hash algorithm to obtain a corresponding first Hash value and a corresponding second Hash value;
the uploading unit is used for adding the first hash value and the second hash value into the same array to generate a first collection; and uploading the first collection to an enterprise underwriting management server for storage.
Further, the information recognition apparatus further includes:
the second acquisition unit is used for acquiring the underwriting data of the historical insurance policy of the user based on the spark data calculation engine;
the third calculation unit is used for calculating the user historical insurance policy underwriting data to obtain the underwriting years of the user;
the sorting unit is used for sorting the underwriting years based on an array sort algorithm to obtain an underwriting year sequence which is arranged from big to small;
the fourth calculation unit is used for sequentially calculating the difference values between the adjacent underwriting years in the underwriting year sequence and sequentially forming the difference values into a difference value sequence;
the fifth calculating unit is used for correspondingly inputting the numerical values in the difference value sequence into a preset formula to calculate the continuous underwriting years of the user;
and the storage unit is used for storing the continuous underwriting years of the user and the user information of the user into a preset database in an associated manner.
Further, the information recognition apparatus further includes:
a third acquiring unit, configured to acquire the number of years of underwriting of the user according to the underwriting years of the user;
a fourth acquiring unit, configured to acquire a year span of underwriting by the user according to a maximum year and a minimum year in the underwriting years of the user;
a sixth calculating unit, configured to calculate a continuous underwriting probability of the user according to the number of years underwriting by the user, the continuous underwriting years, and the year span;
the classification result comprises an underwriting label and a corresponding prediction probability, and the information identification device further comprises:
and the seventh calculating unit is used for carrying out weighted calculation according to the continuous underwriting probability and the prediction probability corresponding to the underwriting label to obtain the underwriting probability of the user.
The present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.
The present application also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, carries out the steps of the method of any of the above.
The information identification method, the information identification device, the computer equipment and the storage medium receive the voice information returned by the target user, and perform voice conversion on the voice information to obtain corresponding user text information; inputting the user text information into a preset recognition model respectively; the recognition model is obtained by fusing a CNN model and a DBM model; extracting local feature vectors of the user text information by a full connection layer of the CNN model, and extracting global feature vectors of the user text information through the DBM model; fusing the local feature vector and the global feature vector to obtain a fused feature vector and inputting the fused feature vector to the classification layer; and finally, carrying out classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, wherein the classification result is used as a recognition result of the voice information returned by the target user. According to the method and the device, local features in the user text information are considered, and global features are combined, so that the final classification result is more accurate.
Drawings
FIG. 1 is a schematic diagram illustrating steps of an information recognition method according to an embodiment of the present application;
FIG. 2 is a block diagram of an information recognition apparatus according to an embodiment of the present application;
fig. 3 is a block diagram illustrating a structure of a computer device according to an embodiment of the present application.
The implementation, functional features and advantages of the objectives of the present application will be further explained with reference to the accompanying drawings.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Referring to fig. 1, an embodiment of the present application provides an information identification method, including the following steps:
step S1, reading the continuous underwriting years of each user stored in the preset database, and selecting a target user with the continuous underwriting years larger than a preset value;
step S2, outputting preset sales text information to the user terminal of the target user;
step S3, receiving the voice information returned by the target user through the user terminal, and performing voice conversion on the voice information to obtain corresponding user text information;
step S4, inputting the user text information into a preset recognition model respectively; the recognition model is obtained by fusing a CNN model and a DBM model, wherein a classification layer which is formed by connecting a full connection layer of the CNN model and a characteristic output layer of the DBM model together is used as a final output layer of the recognition model;
step S5, extracting the local feature vector of the user text information by the full connection layer of the CNN model, and extracting the global feature vector of the user text information by the DBM model;
step S6, fusing the local feature vector and the global feature vector to obtain a fused feature vector and inputting the fused feature vector to the classification layer;
and step S7, performing classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, and using the classification result as a recognition result of the voice information returned by the target user.
In this embodiment, the information identification method is applied to a scene of acquiring the client underwriting requirements, and is implemented based on an intelligent robot, and the intelligent robot replaces manpower to identify the client underwriting requirements, so that not only is the labor cost reduced, but also inaccurate identification caused by artificial subjective consciousness can be avoided.
In this embodiment, the above method is directed to high quality customers. As described in the above step S1, the high-quality client in the present embodiment refers to the target user whose consecutive underwriting years are greater than the preset value; when the target user has more continuous underwriting years, the probability of underwriting of the target user is correspondingly higher, and therefore the target user is defined as a high-quality client.
As described in the step S2, for the target user, the preset marketing text information is output to the user terminal where the target user is located, so that the target user can conveniently view the corresponding marketing text information, where the marketing text information includes related information such as insurance content and benefit information for marketing.
As described in the step S3, the target user may make a corresponding answer to the sales text information, and the target user may answer the sales text information with voice for the convenience of the user, and the user terminal where the target user is located collects the voice information of the target user and then sends the voice information to the intelligent robot; the intelligent robot can convert the voice information into user text information.
As described in step S4, a recognition model (CNN-DBM model) is obtained by pre-training, and the recognition model is a new model trained based on a CNN (convolutional neural network) model and a DBM (data-based mechanical) model, and the recognition model fuses the CNN model and the DBM model, so that the recognition model has a better effect in classifying user text information and has a better recognition capability for professional words and texts. Specifically, a CNN model and a DBM model are fused in the recognition model, and a full connection layer of the CNN model and a feature output layer of the DBM model are connected with a classification layer as a final output layer of the recognition model; that is, the recognition model in this embodiment includes not only the CNN model and the DBM model, but also the CNN model and the DBM model do not output the classification result separately, and the CNN model and the DBM model perform feature extraction, respectively, and output the classification result by the classification layer of the recognition model after fusing the extracted features. In this embodiment, when the recognition model is obtained through training, the CNN model and the DBM model need to be trained respectively, and then the features of the CNN model and the DBM model are extracted respectively and input to the classification layer for fusion training to obtain the recognition model.
As described in step S5, the feature vectors of the user text information are extracted by the CNN model and the DBM model in the recognition model, wherein the CNN model extracts the corresponding feature vectors from the last layer (i.e., the layer before the output layer) of the full connection layer, and the extracted feature vectors are local feature vectors because the CNN model extracts features by sliding the convolution kernel when extracting the feature vectors. The DBM model is an undirected graph connected neural network, and text representation is more accurate through joint sampling calculation of two connected layers of nodes, so that the DBM model extracts global feature vectors, and the DBM model extracts the global feature vectors through an output layer of the DBM model. In the CNN model, the last layer of the full-link layer outputs local feature vectors corresponding to 2 output categories (classification results include two labels, an underwriting label and a non-underwriting label). The DBM model also corresponds to 2 output categories; the CNN model and the DBM model extract corresponding feature vectors from each output category, and then the local feature vectors and the global feature vectors of each output category are fused.
As described in step S6, the local feature vector and the global feature vector are fused, and a corresponding fused feature vector is output. The CNN model and the DBM model extract different feature vectors of the user text information, and the target feature vector obtained after fusion has different characteristics of the feature vectors.
As described in step S7 above, in the present embodiment, the fused feature vector is input to the classification layer and calculated, and the principle of calculation is the same as that in the case of input of a single feature vector, but the input vector (here, the fused feature vector obtained by fusion) is different. The fusion feature vector formed by fusing the two feature vectors is a new vector, the new vector is input into a classification layer for calculation and finally classified, the characteristics of different feature vectors are fused in the classification result, and the obtained classification result is more accurate than the identification classification when the feature vectors are independently input. Therefore, the effect of identifying and classifying the underwriting requirements of the user text information is better, and the real underwriting requirements of the user can be identified more accurately.
In an embodiment, after the step S7 of performing classification calculation on the fusion feature vector through the classification layer to obtain a classification result corresponding to the user text information, and taking the classification result as a recognition result of the speech information returned by the target user, the method includes:
step S8, creating a blank layer; the blank layer is used for adding the sales text information, the user text information and the classification result, so that the information generated in the process is conveniently solidified, and the subsequent survey is conveniently carried out.
Step S9, acquiring the sales text information and the character number of the user text information; the number of characters is the number of characters in the text information.
Step S10, calculating a first area size and a second area size occupied by the sales text information and the user text information added to the blank layer according to the sales text information, the number of characters of the user text information, and a preset font format. The larger the number of characters, the larger the font size, the larger the area it occupies. In this embodiment, the sales text information and the user text information are added to the blank layer in the same preset font format.
Step S11, dividing a first area corresponding to the size of the first area and a second area corresponding to the size of the second area in the blank layer; and dividing a third area with a preset size between the first area and the second area in the blank layer.
Step S12, adding the sales text information to the first area, adding the user text information to the second area, adding the classification result to the third area, and synthesizing into a picture.
In this embodiment, the sales text information and the user text information are respectively added to different areas of the blank layer, so as to distinguish the information conveniently; meanwhile, the classification result is added in the blank layer, so that the classification result can be directly obtained when the subsequent checking is facilitated. In the embodiment, the text information is solidified in the same layer to form the picture, so that the information is conveniently recorded, and the messy data is avoided; and data loss and tampering can be avoided.
In an embodiment, the step S12 of adding the sales text information to the first area, adding the user text information to the second area, and adding the classification result to the third area includes:
step S121, performing hash calculation on the sales text information, the user text information and the classification result respectively to obtain a corresponding sales hash value, a corresponding user hash value and a corresponding classification hash value;
step S122, encrypting the sales text information through the classified hash value, and adding the encrypted sales text information into the first area;
step S123, encrypting the user text information through the sales hash value, and adding the encrypted user text information into the second area;
step S124, encrypt the classification result according to the user hash value, and add the encrypted classification result to the third area.
In this embodiment, the sales text information, the user text information, and the classification result usually include sensitive information of the user, and in order to avoid leakage of the sensitive information, the sales text information, the user text information, and the classification result need to be encrypted.
In this embodiment, the encryption method is that hash calculation is respectively performed on the sales text information, the user text information, and the classification result to obtain a corresponding sales hash value, a user hash value, and a classification hash value; the sales text information is encrypted through the classified Hash values, the user text information is encrypted through the sales Hash values, the classification results are encrypted through the user Hash values, the sales text information, the user text information and the classification results are encrypted through the ring-and-ring-buckled encryption mode, and the method has the following advantages: firstly, the encryption mode is novel, and the encryption of the next information needs to be combined with the hash value characteristic of the previous information; secondly, when the passwords are stored, only any one of the encrypted passwords needs to be stored, so that the difficulty of memorizing the passwords is reduced; after one of the information is decrypted by the encryption password, all the information can be acquired in sequence. When other people do not know the arrangement sequence of the information and the decryption rule, even if other people know one of the encryption passwords, the other people cannot continuously unlock the password of the next information.
In one embodiment, the classification result includes an underwriting label and a corresponding prediction probability; after step S7 of performing classification calculation on the fusion feature vector through the classification layer to obtain a classification result corresponding to the user text information, and taking the classification result as a recognition result of the speech information returned by the target user, the method includes:
step S71, storing the sales text information and the user text information in the preset database;
step S72, combining the sales text information and the underwriting label to obtain a first combination;
step S73, combining the user text information and the prediction probability corresponding to the underwriting label to obtain a second combination;
step S74, respectively calculating the first combination and the second combination through a Hash algorithm to obtain a corresponding first Hash value and a second Hash value;
step S75, adding the first hash value and the second hash value into the same array to generate a first collection; and uploading the first collection to an enterprise underwriting management server for storage.
In this embodiment, in order to facilitate the subsequent review and investigation of the sales text information and the user text information and to avoid the user tampering with the sales text information and the user text information, the sales text information and the user text information need to be stored in the enterprise underwriting management server; however, if the sales text information and the user text information are both stored in the enterprise underwriting management server, the storage cost of the enterprise underwriting management server will be increased, and the pressure of the server will be increased.
Therefore, in this embodiment, the sales text information and the user text information are only stored in a local preset database, so that the storage pressure of the enterprise underwriting management server can be reduced; in order to prevent the sales text information and the user text information from being tampered, in this embodiment, the sales text information and the underwriting label are combined to obtain a first combination, and the user text information and the prediction probability corresponding to the underwriting label are combined to obtain a second combination; then, respectively calculating the first combination and the second combination through a Hash algorithm to obtain a corresponding first Hash value and a corresponding second Hash value; finally, packing the first hash value and the second hash value into a first collection; uploading the first collection to an enterprise underwriting management server for storage; the enterprise underwriting management server only stores one collection, and the collection does not increase too much storage pressure; the collection has the advantages that when the locally stored sales text information and the user text information need to be surveyed, whether the information is real or not is verified, and only a second collection needs to be regenerated according to the method; comparing the second collection with the first collection stored in the enterprise underwriting management server, and if the second collection is consistent with the first collection, indicating that the information is not tampered and is real information; if not, the information is tampered.
In an embodiment, before the step S1 of reading the consecutive underwriting years of each user stored in the preset database and selecting the target user with the consecutive underwriting years greater than the preset value, the method includes:
step S101, acquiring historical insurance policy underwriting data of a user based on a spark data calculation engine;
step S102, calculating the user historical insurance policy underwriting data to obtain the underwriting years of the user;
s103, sequencing the underwriting years based on an array sort algorithm to obtain an underwriting year sequence which is arranged from big to small;
step S104, calculating the difference values between the adjacent underwriting years in the underwriting year sequence in sequence, and forming the difference values into a difference value sequence in sequence;
step S105, correspondingly inputting the numerical values in the difference sequence into a preset formula to calculate the continuous underwriting years of the user;
and step S106, storing the continuous underwriting years of the user and the user information of the user into a preset database in a correlated manner.
In step S101, the historical insurance policy underwriting data of the user includes insurance policy data related to underwriting of the user during the application period, such as historical underwriting years or specific underwriting time. Specifically, the user historical insurance policy underwriting data can be acquired from a background database of the insurance system, the user historical insurance policy underwriting data can be acquired from an insurance webpage through a crawler technology, and the user historical insurance policy underwriting data can be acquired through a data acquisition interface or a big data platform. Preferably, in this embodiment, the Spark calculation engine of the big data platform is used to obtain the historical insurance policy underwriting data of the user, and since the historical insurance policy underwriting data of the user contains massive data, the data obtained by the big data platform is more convenient and efficient, so that the historical insurance policy underwriting data of the user can be further processed subsequently. Spark is an open source cluster computing environment, which enables a memory distributed data set, and can optimize the iterative workload in addition to providing interactive query.
Specifically, all the underwriting policies corresponding to the vehicle IDs are inquired from the user historical insurance policy insurance data, and understandably, due to the fact that the user historical insurance policy insurance data are massive, the calculation amount of the continuous underwriting years is large, and the calculation is repeated, so that the continuous underwriting years can be calculated quickly and accurately through iterative optimization calculation of a big data calculation engine spark.
In the step S102, the underwriting years refer to corresponding years in the closing time of the insurance policy, and if the user a underwrits in 2014, 2016, 2017, 2018, and 2019, the underwriting years are 2014, 2016, 2017, 2018, and 2019.
In step S103, the underwriting year sequence is a set of data in which the numerical values of underwriting years are arranged in a predetermined order as elements. If the underwriting years of the vehicle C are 2012, 2013, 2016 and 2018, respectively, the sequence of underwriting years may be 2012, 2013, 2016 and 2018, or 2018, 2016, 2013 and 2012. Specifically, in this embodiment, the underwriting year sequences are obtained by sorting the underwriting year values in descending order or descending order through an array sort method.
As the user may not be a continuous underwriting, for example, the years of underwriting are 2012, 2013, 2015, 2016; at this time, although the user has not been underwritten in 2014, the continuous underwriting years of the user in the general sense are only two years, obviously, the continuous underwriting probability of the user is high, and the calculation should not be performed according to two years, so that the comprehensive calculation can be performed by combining the years of the continuous underwriting of the user for many times to obtain a comprehensive continuous underwriting year; or the user's underwriting years are 2012, 2014, 2016, 2018, and an equivalent consecutive underwriting year can be calculated.
As described in the above steps S104 and S105, in this embodiment, the difference between adjacent underwriting years in the sequence of underwriting years is sequentially calculated, and if the difference is 1, it indicates that the two adjacent underwriting years are consecutive underwriting years; if the difference is greater than 1, it indicates that there is an unapproved year between the two adjacent years, but the user still has underwriting behavior in the two adjacent years. At this time, the numerical values in the difference value sequence are correspondingly input into a preset formula to calculate the continuous underwriting years of the user, which is an equivalent continuous underwriting years, and obviously, the equivalent continuous underwriting years are lower when the difference value is larger.
In one embodiment, the difference sequence formed by the differences in sequence is { a }1,a2,a3,…an}，
The preset formula is a formula I:
in this embodiment, the difference values may be arranged from small to large, and the arranged difference values are sequentially combined into the difference value sequence { a }1,a2,a3,…an}; this is because the smaller the difference value is, the greater the influence of the difference value on the probability of continuous underwriting is, and the more the probability of underwriting of the user is represented, so that the smaller the difference value is ranked ahead, and the greater the influence of the difference value on the final calculation result is in the calculation; if the difference value is larger, the larger the number of years indicating that the difference value is not continuously underwritten, the smaller the influence of the difference value on the probability of continuously underwriting is, so that the influence of the difference value after being arranged is smaller on the final calculation result in the calculation.
In another embodiment, the preset formula for calculating the continuous underwriting years of the user is formula two:
in the present embodiment, since the underwriting year sequence is arranged from large to small, it can be understood that the more recent years (e.g., 2018, 2017) are, the more likely the underwriting of the user is represented, the more referential the underwriting year data of the user is; while the relative availability of the underwriting years farther from the current time is lower. Therefore, the further back in the calculation process in the above formula (i.e., the farther the underwriting year is from the current year), the less the impact it has on the results of the calculation of the number of consecutive underwriting years by the end user.
The series of underwriting years is exemplified by {2018, 2017, 2015, 2010}, and in the conventional definition of consecutive underwriting years, the consecutive underwriting years are 2, since only 2018 and 2017 are consecutive underwriting years. In this embodiment, the differences between adjacent underwriting years in the underwriting year sequence are sequentially calculated to be 1, 2, 5, and the differences are sequentially combined to form a difference sequence {1, 2, 5 }; and then correspondingly inputting the numerical values in the difference sequence into a formula II, and calculating to obtain the continuous underwriting years as follows:
the calculated successive underwriting years are obviously greater than the usual successive underwriting years (i.e. 2), and the calculation result obviously reflects the real underwriting requirement of the user more truly.
In another embodiment, after the step S105 of inputting the numerical value in the difference value sequence into a preset formula to calculate the number of consecutive underwriting years of the user, the method includes:
step S107, acquiring the years of the user underwriting according to the underwriting years of the user;
step S108, acquiring the year span of underwriting of the user according to the maximum year and the minimum year in the underwriting years of the user;
step S109, calculating the continuous underwriting probability of the user according to the years of underwriting of the user, the continuous underwriting years and the year span;
the number of years of the user underwriting is p, the number of continuous underwriting years is q, the year span is m, the equivalent continuous underwriting probability is W, and the calculation formula of the continuous underwriting probability is as follows:
it will be appreciated that the continuous underwriting probability of the user in the current year is closely related to the years previously underwritten by the user, the span of the underwriting years, and the continuous underwriting years, and therefore an equivalent continuous underwriting probability of the user in the current year can be calculated from the above parameters.
Taking the series of underwriting years {2018, 2017, 2015, 2010} in the above embodiment as an example, wherein the number of consecutive underwriting years q obtained by the above calculation process is 73/24, the number of years p underwriting by the user is 4, the span m of years is 9, and the equivalent consecutive underwriting probability is given as the above example
I.e. the user has a current equivalent continuous underwriting probability of 15%. Further, according to the corresponding relation between the equivalent continuous underwriting probability and the sales terminal, the sales terminal corresponding to the equivalent continuous underwriting probability is determined; and acquiring the user information of the user and the terminal information of the sales terminal, sequentially synthesizing the user information of the user and the terminal information of the sales terminal into a preset image layer, generating a synthetic picture, sending the synthetic picture to the sales terminal, and storing the synthetic picture in a preset database. The information is generated into a picture, so that the information is conveniently solidified in the same picture, the information is conveniently checked, and the missing and losing of the information are avoided; and the data is stored in a database, so that subsequent tracking and query are facilitated.
In this embodiment, the classifying result includes an underwriting label and a corresponding prediction probability, and after the step S7 of performing classification calculation on the fusion feature vector through the classification layer to obtain the classifying result corresponding to the user text information, and taking the classifying result as the recognition result of the voice information returned by the target user, the method includes:
and step S8a, carrying out weighted calculation according to the continuous underwriting probability and the prediction probability corresponding to the underwriting label to obtain the underwriting probability of the user. In the embodiment, the continuous underwriting probability calculated by the method and the prediction probability predicted by the recognition model are combined to be calculated comprehensively, so that the final result has higher referential property and can reflect the real underwriting requirement of the user.
Referring to fig. 2, an embodiment of the present application further provides an information identification apparatus, including:
the selection unit 10 is used for reading the continuous underwriting years of each user stored in the preset database and selecting a target user with the continuous underwriting years larger than a preset value from the continuous underwriting years;
the output unit 20 is configured to output preset sales text information to a user terminal where the target user is located;
a conversion unit 30, configured to receive voice information returned by the target user through the user terminal, and perform voice conversion on the voice information to obtain corresponding user text information;
the input unit 40 is used for inputting the user text information into preset recognition models respectively; the recognition model is obtained by fusing a CNN model and a DBM model, wherein a classification layer which is formed by connecting a full connection layer of the CNN model and a characteristic output layer of the DBM model together is used as a final output layer of the recognition model;
an extracting unit 50, configured to extract a local feature vector of the user text information from the fully connected layer of the CNN model, and extract a global feature vector of the user text information through the DBM model;
a fusion unit 60, configured to fuse the local feature vector and the global feature vector to obtain a fusion feature vector, and input the fusion feature vector to the classification layer;
and the classifying unit 70 is configured to perform classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, and the classification result is used as a recognition result of the voice information returned by the target user.
In an embodiment, the information identification apparatus further includes:
the creating unit is used for creating a blank layer;
a first acquiring unit, configured to acquire the sales text information and the number of characters of the user text information;
the first calculation unit is used for calculating the size of a first area and the size of a second area correspondingly occupied by the sales text information and the user text information added into the blank layer according to the sales text information, the number of characters of the user text information and a preset font format;
the dividing unit is used for dividing a first area corresponding to the size of the first area and a second area corresponding to the size of the second area in the blank layer; dividing a third area with a preset size between the first area and the second area in the blank layer;
and the adding unit is used for adding the sales text information into the first area, adding the user text information into the second area, adding the classification result into the third area and synthesizing the classification result into a picture.
In one embodiment, the adding unit includes:
the calculation subunit is configured to perform hash calculation on the sales text information, the user text information, and the classification result, respectively, to obtain a corresponding sales hash value, a corresponding user hash value, and a corresponding classification hash value;
the first adding subunit is used for encrypting the sales text information through the classified hash value and adding the encrypted sales text information into the first area;
the second adding subunit is used for encrypting the user text information through the sales hash value and adding the encrypted user text information into the second area;
and the third adding subunit is configured to encrypt the classification result according to the user hash value, and add the encrypted classification result to the third area.
In one embodiment, the classification result includes an underwriting label and a corresponding prediction probability; the information recognition apparatus further includes:
the storage unit is used for storing the sales text information and the user text information in the preset database;
the first combination unit is used for combining the sales text information and the underwriting label to obtain a first combination;
the second combination unit is used for combining the user text information with the prediction probability corresponding to the underwriting label to obtain a second combination;
the second calculation unit is used for calculating the first combination and the second combination respectively through a Hash algorithm to obtain a corresponding first Hash value and a corresponding second Hash value;
the uploading unit is used for adding the first hash value and the second hash value into the same array to generate a first collection; and uploading the first collection to an enterprise underwriting management server for storage.
In an embodiment, the information identification apparatus further includes:
the second acquisition unit is used for acquiring the underwriting data of the historical insurance policy of the user based on the spark data calculation engine;
the third calculation unit is used for calculating the user historical insurance policy underwriting data to obtain the underwriting years of the user;
the sorting unit is used for sorting the underwriting years based on an array sort algorithm to obtain an underwriting year sequence which is arranged from big to small;
the fourth calculation unit is used for sequentially calculating the difference values between the adjacent underwriting years in the underwriting year sequence and sequentially forming the difference values into a difference value sequence;
the fifth calculating unit is used for correspondingly inputting the numerical values in the difference value sequence into a preset formula to calculate the continuous underwriting years of the user;
and the storage unit is used for storing the continuous underwriting years of the user and the user information of the user into a preset database in an associated manner.
In an embodiment, the information identification apparatus further includes:
a third acquiring unit, configured to acquire the number of years of underwriting of the user according to the underwriting years of the user;
a fourth acquiring unit, configured to acquire a year span of underwriting by the user according to a maximum year and a minimum year in the underwriting years of the user;
a sixth calculating unit, configured to calculate a continuous underwriting probability of the user according to the number of years underwriting by the user, the continuous underwriting years, and the year span;
the classification result comprises an underwriting label and a corresponding prediction probability, and the information identification device further comprises:
and the seventh calculating unit is used for carrying out weighted calculation according to the continuous underwriting probability and the prediction probability corresponding to the underwriting label to obtain the underwriting probability of the user.
In this embodiment, please refer to the corresponding description in the method embodiment for the specific implementation of the units/sub-units, which is not described herein again.
Referring to fig. 3, a computer device, which may be a server and whose internal structure may be as shown in fig. 3, is also provided in the embodiment of the present application. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the computer designed processor is used to provide computational and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing text information and the like. The network interface of the computer device is used for communicating with an external terminal through a network connection. When executed by the processor, the computer program implements the information identification method in the above method embodiments, and will not be described repeatedly herein.
Those skilled in the art will appreciate that the architecture shown in fig. 3 is only a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects may be applied.
An embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the information identification method in the foregoing method embodiment is implemented, and repeated descriptions are not repeated here. It is to be understood that the computer-readable storage medium in the present embodiment may be a volatile-readable storage medium or a non-volatile-readable storage medium.
In summary, the information identification method, apparatus, computer device and storage medium provided in the embodiments of the present application receive the voice information returned by the target user, and perform voice conversion on the voice information to obtain corresponding user text information; inputting the user text information into a preset recognition model respectively; the recognition model is obtained by fusing a CNN model and a DBM model; extracting local feature vectors of the user text information by a full connection layer of the CNN model, and extracting global feature vectors of the user text information through the DBM model; fusing the local feature vector and the global feature vector to obtain a fused feature vector and inputting the fused feature vector to the classification layer; and finally, carrying out classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, wherein the classification result is used as a recognition result of the voice information returned by the target user. According to the method and the device, local features in the user text information are considered, and global features are combined, so that the final classification result is more accurate.
It will be understood by those of ordinary skill in the art that all or a portion of the processes of the methods of the embodiments described above may be implemented by a computer program that may be stored on a non-volatile computer-readable storage medium, which when executed, may include the processes of the embodiments of the methods described above, wherein any reference to memory, storage, database, or other medium provided herein and used in the embodiments may include non-volatile and/or volatile memory.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that includes the element.
The above description is only for the preferred embodiment of the present application and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are intended to be included within the scope of the present application.
Claims (10)
1. An information identification method, characterized by comprising the steps of:
reading the continuous underwriting years of each user stored in a preset database, and selecting a target user with the continuous underwriting years larger than a preset value;
outputting preset sales text information to a user terminal where the target user is located;
receiving voice information returned by the target user through the user terminal, and carrying out voice conversion on the voice information to obtain corresponding user text information;
inputting the user text information into a preset recognition model respectively; the recognition model is obtained by fusing a CNN model and a DBM model, wherein a classification layer which is formed by connecting a full connection layer of the CNN model and a characteristic output layer of the DBM model together is used as a final output layer of the recognition model;
extracting local feature vectors of the user text information by a full connection layer of the CNN model, and extracting global feature vectors of the user text information through the DBM model;
fusing the local feature vector and the global feature vector to obtain a fused feature vector and inputting the fused feature vector to the classification layer;
and carrying out classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, wherein the classification result is used as a recognition result of the voice information returned by the target user.
2. The information recognition method according to claim 1, wherein after the step of performing classification calculation on the fusion feature vector through the classification layer to obtain a classification result corresponding to the user text information as a recognition result of the voice information returned by the target user, the method comprises:
creating a blank layer;
acquiring the sales text information and the number of characters of the user text information;
calculating the size of a first area and the size of a second area correspondingly occupied by the sales text information and the user text information added into the blank layer according to the sales text information, the number of characters of the user text information and a preset font format;
dividing a first area corresponding to the size of the first area and a second area corresponding to the size of the second area in the blank layer; dividing a third area with a preset size between the first area and the second area in the blank layer;
adding the sales text information into the first area, adding the user text information into the second area, adding the classification result into the third area, and synthesizing into a picture.
3. The information recognition method according to claim 2, wherein the step of adding the sales text information to the first area, the user text information to the second area, and the classification result to the third area includes:
performing hash calculation on the sales text information, the user text information and the classification result respectively to obtain a corresponding sales hash value, a corresponding user hash value and a corresponding classification hash value;
encrypting the sales text information through the classified hash value, and adding the encrypted sales text information into the first area;
encrypting the user text information through the sales hash value, and adding the encrypted user text information into the second area;
and encrypting the classification result through the user hash value, and adding the encrypted classification result into the third area.
4. The information identification method according to claim 1, wherein the classification result comprises an underwriting label and a corresponding prediction probability; after the step of performing classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, and taking the classification result as a recognition result of the voice information returned by the target user, the method includes:
saving the sales text information and the user text information in the preset database;
combining the sales text information with the underwriting label to obtain a first combination;
combining the user text information with the prediction probability corresponding to the underwriting label to obtain a second combination;
respectively calculating the first combination and the second combination through a Hash algorithm to obtain a corresponding first Hash value and a corresponding second Hash value;
adding the first hash value and the second hash value into the same array to generate a first collection; and uploading the first collection to an enterprise underwriting management server for storage.
5. The information identification method according to claim 1, wherein the step of reading the continuous underwriting years of each user stored in the preset database and selecting the target user with the continuous underwriting years larger than the preset value comprises:
acquiring historical insurance policy underwriting data of a user based on a spark data calculation engine;
calculating the historical insurance policy underwriting data of the user to obtain the underwriting years of the user;
sequencing the underwriting years based on an array sort algorithm to obtain an underwriting year sequence which is arranged from big to small;
sequentially calculating differences between adjacent underwriting years in the underwriting year sequence, and sequentially forming the differences into a difference sequence;
correspondingly inputting the numerical values in the difference value sequence into a preset formula to calculate the continuous underwriting years of the user;
and storing the continuous underwriting years of the user and the user information of the user into a preset database in a correlation manner.
6. The information identification method according to claim 5, wherein the step of inputting the numerical value correspondence in the difference value sequence into a preset formula to calculate the continuous underwriting years of the user is followed by:
acquiring the years of the user underwriting according to the underwriting years of the user;
acquiring the year span of the underwriting of the user according to the maximum year and the minimum year in the underwriting years of the user;
calculating the continuous underwriting probability of the user according to the years of underwriting of the user, the continuous underwriting years and the year span;
the classification result comprises an underwriting label and a corresponding prediction probability, and after the step of performing classification calculation on the fusion feature vector through the classification layer to obtain a classification result corresponding to the user text information as a recognition result of the voice information returned by the target user, the method comprises the following steps of:
and carrying out weighted calculation according to the continuous underwriting probability and the prediction probability corresponding to the underwriting label to obtain the underwriting probability of the user.
7. An information identifying apparatus, comprising:
the selection unit is used for reading the continuous underwriting years of each user stored in the preset database and selecting a target user with the continuous underwriting years larger than a preset value from the continuous underwriting years;
the output unit is used for outputting preset sales text information to a user terminal where the target user is located;
the conversion unit is used for receiving voice information returned by the target user through the user terminal and carrying out voice conversion on the voice information to obtain corresponding user text information;
the input unit is used for respectively inputting the user text information into a preset recognition model; the recognition model is obtained by fusing a CNN model and a DBM model, wherein a classification layer which is formed by connecting a full connection layer of the CNN model and a characteristic output layer of the DBM model together is used as a final output layer of the recognition model;
the extraction unit is used for extracting the local feature vector of the user text information from the full connection layer of the CNN model and extracting the global feature vector of the user text information through the DBM model;
the fusion unit is used for fusing the local feature vector and the global feature vector to obtain a fusion feature vector and inputting the fusion feature vector to the classification layer;
and the classification unit is used for performing classification calculation on the fusion feature vectors through the classification layer to obtain a classification result corresponding to the user text information, and the classification result is used as a recognition result of the voice information returned by the target user.
8. The information identifying apparatus according to claim 7, further comprising:
the creating unit is used for creating a blank layer;
a first acquiring unit, configured to acquire the sales text information and the number of characters of the user text information;
the first calculation unit is used for calculating the size of a first area and the size of a second area correspondingly occupied by the sales text information and the user text information added into the blank layer according to the sales text information, the number of characters of the user text information and a preset font format;
the dividing unit is used for dividing a first area corresponding to the size of the first area and a second area corresponding to the size of the second area in the blank layer; dividing a third area with a preset size between the first area and the second area in the blank layer;
and the adding unit is used for adding the sales text information into the first area, adding the user text information into the second area, adding the classification result into the third area and synthesizing the classification result into a picture.
9. A computer device comprising a memory and a processor, the memory having stored therein a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method according to any of claims 1 to 6.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010124754.3A CN111414451A (en) | 2020-02-27 | 2020-02-27 | Information identification method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010124754.3A CN111414451A (en) | 2020-02-27 | 2020-02-27 | Information identification method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111414451A true CN111414451A (en) | 2020-07-14 |
Family
ID=71492736
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010124754.3A Pending CN111414451A (en) | 2020-02-27 | 2020-02-27 | Information identification method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111414451A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112632222A (en) * | 2020-12-25 | 2021-04-09 | 海信视像科技股份有限公司 | Terminal equipment and method for determining data belonging field |
| CN112749278A (en) * | 2020-12-30 | 2021-05-04 | 华南理工大学 | Classification method for building engineering change instructions |
| CN112818859A (en) * | 2021-02-02 | 2021-05-18 | 电子科技大学 | Deep hash-based multi-level retrieval pedestrian re-identification method |
-
2020
- 2020-02-27 CN CN202010124754.3A patent/CN111414451A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112632222A (en) * | 2020-12-25 | 2021-04-09 | 海信视像科技股份有限公司 | Terminal equipment and method for determining data belonging field |
| CN112632222B (en) * | 2020-12-25 | 2023-02-03 | 海信视像科技股份有限公司 | Terminal equipment and method for determining data belonging field |
| CN112749278A (en) * | 2020-12-30 | 2021-05-04 | 华南理工大学 | Classification method for building engineering change instructions |
| CN112818859A (en) * | 2021-02-02 | 2021-05-18 | 电子科技大学 | Deep hash-based multi-level retrieval pedestrian re-identification method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9218364B1 (en) | Monitoring an any-image labeling engine | |
| CN111414451A (en) | Information identification method and device, computer equipment and storage medium | |
| KR20180118596A (en) | Semi-automatic labeling of data sets | |
| CN114201621A (en) | Cross-modal retrieval model construction and retrieval method based on image-text cooperative attention | |
| CN110781460A (en) | Copyright authentication method, device, equipment, system and computer readable storage medium | |
| CN111582409A (en) | Training method of image label classification network, image label classification method and device | |
| CN113377936B (en) | Intelligent question and answer method, device and equipment | |
| Okon et al. | An improved online book recommender system using collaborative filtering algorithm | |
| US11966829B2 (en) | Convolutional artificial neural network based recognition system in which registration, search, and reproduction of image and video are divided between and performed by mobile device and server | |
| CN113435330B (en) | Video-based micro-expression recognition method, device, equipment and storage medium | |
| CN113378710A (en) | Layout analysis method and device for image file, computer equipment and storage medium | |
| CN113873330B (en) | Video recommendation method and device, computer equipment and storage medium | |
| CN112232971A (en) | Anti-fraud detection method, anti-fraud detection device, computer equipment and storage medium | |
| CN112288279A (en) | Business risk assessment method and device based on natural language processing and linear regression | |
| CN110705490A (en) | Visual emotion recognition method | |
| CN110213660B (en) | Program distribution method, system, computer device and storage medium | |
| CN113537206A (en) | Pushed data detection method and device, computer equipment and storage medium | |
| CN116501979A (en) | Information recommendation method, information recommendation device, computer equipment and computer readable storage medium | |
| CN113497978B (en) | Video scene classification method, device, server and storage medium | |
| Thonglek et al. | Ivaa: Intelligent vehicle accident analysis system | |
| CN115937556A (en) | Object identification method, device, equipment and storage medium | |
| CN113888216A (en) | Advertisement information pushing method and device, electronic equipment and storage medium | |
| CN112115443A (en) | Terminal user authentication method and system | |
| Thonglek et al. | Vehicle part damage analysis platform for autoinsurance application | |
| CN113807870B (en) | Vehicle information authentication method, device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |