CN114925920B - Offline position prediction method and device, electronic equipment and storage medium - Google Patents

Offline position prediction method and device, electronic equipment and storage medium Download PDF

Info

Publication number
CN114925920B
CN114925920B CN202210583200.9A CN202210583200A CN114925920B CN 114925920 B CN114925920 B CN 114925920B CN 202210583200 A CN202210583200 A CN 202210583200A CN 114925920 B CN114925920 B CN 114925920B
Authority
CN
China
Prior art keywords
sample set
sample
information
user
position prediction
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.)
Active
Application number
CN202210583200.9A
Other languages
Chinese (zh)
Other versions
CN114925920A (en
Inventor
唐隆伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Property and Casualty Insurance Company of China Ltd
Original Assignee
Ping An Property and Casualty Insurance Company of China Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ping An Property and Casualty Insurance Company of China Ltd filed Critical Ping An Property and Casualty Insurance Company of China Ltd
Priority to CN202210583200.9A priority Critical patent/CN114925920B/en
Publication of CN114925920A publication Critical patent/CN114925920A/en
Application granted granted Critical
Publication of CN114925920B publication Critical patent/CN114925920B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • 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
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • 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
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Quality & Reliability (AREA)
  • General Business, Economics & Management (AREA)
  • Medical Informatics (AREA)
  • Operations Research (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Tourism & Hospitality (AREA)
  • Remote Sensing (AREA)
  • Probability & Statistics with Applications (AREA)
  • Game Theory and Decision Science (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to an intelligent decision-making technology, and discloses an offline position prediction method, which comprises the following steps: carrying out normalization processing on position information and time information in a sample set, and eliminating invalid features in the sample set; respectively carrying out predictive training on a preset number of position predictive models by using a sample set after invalid features are removed until the predictive training meets preset conditions, and exiting the predictive training; calculating to obtain a reconciliation value between the precision rate and the recall rate of each position prediction model by using the prediction result and the real result, and selecting the position prediction model with the highest reconciliation value as a target position prediction model; and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by using the target position prediction model. The invention also provides an offline position prediction device, equipment and a medium. The invention can solve the problem that the user cannot acquire the geographical position information of the user when the user is in an off-line state.

Description

Offline position prediction method and device, electronic equipment and storage medium
Technical Field
The present invention relates to the field of intelligent decision making technologies, and in particular, to an offline position prediction method, an offline position prediction device, an electronic device, and a computer readable storage medium.
Background
Providing differentiated services based on location information of users is an important means of public welfare operation and business operation. How to acquire the position information of the user is currently common to deploy a service APP at a mobile terminal of the user, and to build a position service function in the service APP, and under the premise of user authorization, acquire the GPS information of the user in real time through the position service function, so as to identify the position information of the user according to the GPS information of the user.
The method depends on the fact that a user opens the service APP and the user authorizes the service APP to use the GPS positioning function, and when the user does not open the service APP or does not use the GPS positioning function, the position information of the user cannot be acquired, and differentiated services based on the position information cannot be provided for the user. If the GPS information of the user is obtained under the premise that the user does not use the service APP or is unauthorized, the privacy of the user can be violated. Therefore, how to obtain the location information of the user is a problem to be solved in the case of offline service APP or offline GPS positioning function.
Disclosure of Invention
The invention provides an offline position prediction method, an offline position prediction device and a computer readable storage medium, and mainly aims to improve the accuracy of offline position prediction.
In order to achieve the above object, the present invention provides an offline position prediction method, including:
acquiring a sample set of user historical activity information, wherein the historical activity information comprises position information and time information corresponding to the position information;
Respectively carrying out normalization processing on the position information and the time information;
Performing invalid characteristic eliminating operation on the normalized sample set;
Respectively carrying out predictive training on a preset number of position predictive models by using a sample set after invalid features are removed, and exiting the predictive training until the predictive training meets preset conditions to obtain a predictive result of each position predictive model;
Calculating a reconciliation value between the precision rate and the recall rate of each position prediction model by using the prediction result and the real result of the sample set, and selecting a position prediction model with the highest reconciliation value as a target position prediction model;
and acquiring historical activity information of the user to be predicted, and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model.
Optionally, the normalizing the position information and the time information respectively includes:
Marking the time information in the sample set by using a preset time tag to obtain marked time information;
calculating the similarity between the position information, and determining the position information corresponding to the similarity larger than a preset similarity threshold value as the same position set;
And randomly selecting one position information from the same position set to replace each position information in the same position set.
Optionally, the calculating the similarity between each piece of location information includes:
identifying longitude and latitude corresponding to each position information in the sample set;
generating GeoHash codes of each piece of position information by utilizing GeoHash algorithm according to the longitude and latitude of each piece of position information;
And calculating the coding similarity between GeoHash codes of each piece of position information, and taking the coding similarity as the similarity between the corresponding pieces of position information.
Optionally, the removing the invalid feature from the normalized sample set includes:
Extracting the position characteristics of the corresponding samples of each user in the sample set one by one;
Sequentially calculating the percentage between the sample number corresponding to each position feature of each user and the total sample number corresponding to the user to obtain the coverage rate of each position feature of the user;
And determining the position features with the coverage rate smaller than a preset coverage rate threshold as invalid features and eliminating the invalid features.
Optionally, the step of performing, by the root, an invalid feature rejection operation on the normalized sample set is replaced by an invalid sample rejection operation on the normalized sample set, and the invalid sample rejection operation on the normalized sample set includes:
respectively pre-training a preset number of position prediction models by using the sample set, and exiting the pre-training until the pre-training meets the preset training times to obtain a pre-training result of each position prediction model;
counting the repetition rate of the pre-training result of each sample in the sample set in each position prediction model;
and taking the sample with the repetition rate smaller than a preset repetition rate threshold as an invalid sample, and removing the invalid sample from the sample set.
Optionally, the predicting training is performed on a preset number of position predicting models by using the sample set after invalid features are removed, until the predicting training meets a preset condition, the predicting training is exited, and a predicting result of each position predicting model is obtained, including:
Performing predictive training on the extraction of the position features of the sample set with invalid features removed by using each position predictive model to obtain a position feature set;
Calculating probability values between each position feature in the position feature set and a preset number of position tags by using a preset activation function, and selecting the position tag with the maximum probability value as a prediction result;
Judging whether an error value between the predicted result and a real result of the sample set after invalid features are removed meets the preset condition or not by using a preset loss function;
if the error value does not meet the preset condition, adjusting the parameter value of each position prediction model, and returning to the step of extracting the position characteristics of the sample set with invalid characteristics removed by using each position prediction model;
and if the error value meets the preset condition, exiting the predictive training.
Optionally, the calculating, by using the prediction result and the real result of the sample set, a harmonic value between the precision rate and the recall rate of each location prediction model includes:
obtaining a predicted result and a real result of each sample in the sample set;
randomly selecting a prediction result as a reference position;
In each position prediction model, respectively counting a first sample number of which the real result and the predicted result in the sample set are both the reference position, a second sample number of which the real result in the sample set of which the predicted result is the reference position is not the reference position, and a third sample number of which the predicted result in the sample set of which the real result is the reference position is not the reference position;
Calculating the precision rate of each position prediction model according to a precision rate calculation formula by using the first sample number and the second sample number;
calculating the recall rate of each position prediction model according to a recall rate calculation formula by using the first sample number and the third sample number;
and calculating the reconciliation value between the precision rate and the recall rate of each position prediction model according to a reconciliation value calculation formula by utilizing the precision rate and the recall rate.
In order to solve the above problems, the present invention also provides an offline position prediction apparatus, the apparatus comprising:
the sample processing module is used for acquiring a sample set of user historical activity information, wherein the historical activity information comprises position information and time information corresponding to the position information, normalization processing is carried out on the position information and the time information respectively, and invalid characteristic eliminating operation is carried out on the normalized sample set;
The model prediction training module is used for respectively carrying out prediction training on a preset number of position prediction models by utilizing a sample set after invalid features are removed, and stopping the prediction training until the prediction training meets preset conditions to obtain a prediction result of each position prediction model;
The target model selection module is used for calculating a reconciliation value between the precision rate and the recall rate of each position prediction model by utilizing the prediction result and the real result of the sample set, and selecting the position prediction model with the highest reconciliation value as a target position prediction model;
And the target model application module is used for acquiring the historical activity information of the user to be predicted, and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model.
In order to solve the above-mentioned problems, the present invention also provides an electronic apparatus including:
a memory storing at least one computer program; and
And the processor executes the program stored in the memory to realize the offline position prediction method.
In order to solve the above-mentioned problems, the present invention also provides a computer-readable storage medium having stored therein at least one computer program that is executed by a processor in an electronic device to implement the above-mentioned offline position prediction method.
According to the method, the sample set containing the historical activity information of the user position information and the time information is obtained, the time information and the position information in the sample set are normalized, invalid features in the sample set are removed, the quality of the sample set is improved, meanwhile, the sample set is utilized to conduct position prediction training on a preset number of position prediction models, a reconciliation value between the precision rate and the recall rate of each position prediction model is obtained through calculation, the position prediction model with the highest reconciliation value is selected to serve as a target position prediction model, different position predictions are conducted on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model, and the problem that the geographic position information of the user cannot be obtained when the user is in an offline state can be effectively solved.
Drawings
FIG. 1 is a flowchart illustrating an offline position prediction method according to an embodiment of the present invention;
FIG. 2 is a functional block diagram of an offline position prediction apparatus according to an embodiment of the present invention;
Fig. 3 is a schematic structural diagram of an electronic device for implementing the offline position prediction method according to an embodiment of the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The embodiment of the application provides an offline position prediction method. The execution subject of the offline position prediction method includes, but is not limited to, at least one of a server, a terminal, and the like, which can be configured to execute the method provided by the embodiment of the application. In other words, the offline position prediction method may be performed by software or hardware installed in a terminal device or a server device, and the software may be a blockchain platform. The server may be an independent server, or may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (ContentDelivery Network, CDN), and basic cloud computing services such as big data and artificial intelligence platforms.
Referring to fig. 1, a flowchart of an offline position prediction method according to an embodiment of the invention is shown. In this embodiment, the offline position prediction method includes:
s1, acquiring a sample set of historical activity information of a user, wherein the historical activity information comprises position information and time information corresponding to the position information;
In the embodiment of the invention, the sample set consists of historical activity information covering all time periods of the whole day in a preset time period of a large number of users. For example, one sample contains the location information of the user a between 10 months 1 day and 10 months 30 days, which appears every day, and the time information corresponding to each location information.
In the embodiment of the invention, the sample set can be obtained by acquiring the geographic position information of the user after the user authorization when the user uses the preset service APP with the built-in position service.
S2, respectively carrying out normalization processing on the position information and the time information;
It will be appreciated that, in the preset time period, the time of the user moving every day may be different, the time information appearing in a plurality of samples of the same user may be very various, the time information appearing in a large number of samples of a plurality of users may be more various, which may result in the sample concentrating time information being more dispersed, and meanwhile, when the user is moving at the same geographic location at different times, the reflected position information may be different, which may result in the dispersion of the position information, for example, the north door activity of the user in the district where the user is located and the activity of the user in the same district south gate, the north door of the district and the district south gate correspond to two pieces of position information, but the geographic information of the user may be practically generalized. In order to reduce redundant, discrete location information and position information in the sample set, normalization of time information and position information in the sample set is required.
In detail, the normalizing the position information and the time information respectively includes: marking the time information in the sample set by using a preset time tag to obtain marked time information; calculating the similarity between the position information, and determining the position information corresponding to the similarity larger than a preset similarity threshold value as the same position set; and randomly selecting one position information from the same position set to replace each position information in the same position set.
In the embodiment of the present invention, the preset time tag divides different time periods in a day into different time categories according to the historical activity rule of the user, for example, 6:30 to 7: the time corresponding time tags between 30 are breakfast, 7:30 to 9: the time labels corresponding to the time between 00 are the early peak of work, 9:00 to 12: the time labels corresponding to the time between 00 are at am, 12:00 to 13: the time label corresponding to the time between 00 is lunch, and so on, the preset time label may further include: noon break, afternoon hours, peak late in work, dinner, night, late night, early morning, etc.
In detail, the calculating the similarity between each of the position information includes: identifying longitude and latitude corresponding to each position information in the sample set; generating GeoHash codes of each piece of position information by utilizing GeoHash algorithm according to the longitude and latitude of each piece of position information; and calculating the coding similarity between GeoHash codes of each piece of position information, and taking the coding similarity as the similarity between the corresponding pieces of position information.
In the embodiment of the invention, the longitude and latitude corresponding to the geographic position of the user can be obtained by utilizing the GPS positioning function of the mobile equipment of the user. The GeoHash algorithm converts the longitude and latitude of the two-dimensional space region into a group of character strings, namely GeoHash codes, and measures the distance between different regions by comparing the coding similarity between GeoHash codes corresponding to the different regions, wherein the higher the coding similarity is, the smaller the distance between the corresponding regions is, and the closer the corresponding geographic positions are.
S3, performing invalid characteristic eliminating operation on the normalized sample set;
It will be appreciated that due to the presence of objective factors, isolated location information, time information, or location information and time information that are no longer repeated may be present in the user's activity trajectory. For example, the user moves home or changes work, the corresponding home address or work address changes, and for example, the user is uncomfortable to visit a hospital at night, and the position information of the hospital is not appeared in the follow-up period. Such isolated or no longer repeated time information or location information is of little value to the user's location prediction and may be understood as an invalid feature. In order to ensure that the relevant model is effectively trained by using the sample set, invalid special effects in the sample set are also required to be removed in the embodiment of the invention.
In detail, the removing operation of invalid features on the normalized sample set includes: extracting the position characteristics of the corresponding samples of each user in the sample set one by one; sequentially calculating the percentage between the sample number corresponding to each position feature of each user and the total sample number corresponding to the user to obtain the coverage rate of each position feature of the user; and determining the position features with the coverage rate smaller than a preset coverage rate threshold as invalid features and eliminating the invalid features.
In the embodiment of the invention, word2vec model, NLP (Natural Language Processing ) model and other models with word vector conversion function can be adopted to respectively convert the time information and the position information corresponding to each sample in the sample set into word vector to obtain the word vector matrix corresponding to each sample, and further, the word vector matrix can be subjected to position feature extraction, wherein the position feature comprises but is not limited to features such as time labels, position keywords and the like.
In the embodiment of the present invention, for example, taking a certain location keyword feature as an example, assume that the total number of samples corresponding to a certain user is 200, where the number of samples including the location keyword feature is 35, and the coverage rate of the location keyword feature for the user is 35/200×100%, that is, 17.5%.
In the embodiment of the present invention, the preset coverage rate threshold may be set according to the actual total sample size of each user.
In the embodiment of the invention, the isolated and discrete features can be identified by calculating the coverage rate of each position feature.
In another optional embodiment of the present invention, the step of performing, by the root, an invalid feature rejection operation on the normalized sample set is replaced by an invalid sample rejection operation on the normalized sample set, and the invalid sample rejection operation on the normalized sample set includes: respectively pre-training a preset number of position prediction models by using the sample set, and exiting the pre-training until the pre-training meets the preset training times to obtain a pre-training result of each position prediction model; counting the repetition rate of the pre-training result of each sample in the sample set in each position prediction model; and taking the sample with the repetition rate smaller than a preset repetition rate threshold as an invalid sample, and removing the invalid sample from the sample set.
In another optional embodiment of the present invention, the preset position prediction model may be a position prediction model respectively constructed based on algorithms such as KNN, decision tree, random forest, and the like.
In order to determine the class of the unknown sample, the KNN (K-Nearest Neighbor) generally uses all the samples of known classes as references, calculates the distance between the unknown sample and all the known samples, selects K known samples with the Nearest distance to the unknown sample from the known samples, and classifies the unknown sample and the class of the K Nearest Neighbor samples into a class according to a majority rule.
The decision tree is a classification algorithm which is based on a tree structure to represent the interrelationship between each feature in the effective feature set and utilizes the tree structure to classify the sample set.
The random forest is a classification algorithm formed by a plurality of decision trees, and the random forest fuses the prediction results of each decision tree to obtain a final prediction result.
In another embodiment of the present invention, assuming that there are four position prediction models, the corresponding same sample corresponds to four pre-training results, and if three of the pre-training results are identical, that is, the predicted position information in the pre-training results is identical, the repetition rate of the pre-training results of the sample is 3/4, that is, 75%. And if all the four pre-training results are different, the repetition rate of the pre-training results of the sample is 0.
The preset repetition rate threshold is a threshold manually set according to the overall quality of the sample set, and theoretically, the higher the sample quality is, the smaller the difference of the pretraining results of the same sample under the position prediction model based on different algorithms is.
When the repetition rate of the pre-training results of the samples is smaller than the preset repetition rate threshold, that is, the variability of the pre-training results of the same sample in each position prediction model is relatively large, the samples may have problems of characteristic dispersion or data disorder, and the samples are not helpful to the rapid convergence of the position prediction model training, so that the samples are regarded as invalid samples.
S4, respectively carrying out prediction training on a preset number of position prediction models by using a sample set with invalid features removed, and exiting the prediction training until the prediction training meets preset conditions to obtain a prediction result of each position prediction model;
In the embodiment of the invention, the preset number of position prediction models refer to neural network models constructed based on different classification algorithms, such as KNN, decision tree, random forest and other algorithms.
It will be appreciated that the location information in the sample set is a variety of, and illustratively, location tags may be provided at home, work, other addresses, and further, may be provided for more refined classification of the other addresses, e.g., hospital addresses, super-business addresses, etc.
In the embodiment of the present invention, the preset condition may be that when an error value between a real result of the sample and a predicted result of the sample reaches a preset error value threshold, the prediction training is exited. In practical application, the preset condition may also be that when the training frequency of each position prediction model reaches a preset training frequency threshold value, the prediction training is exited.
In detail, the predicting training is performed on a preset number of position predicting models by using the sample set after invalid features are removed, until the predicting training meets preset conditions, the predicting training is exited, and a predicting result of each position predicting model is obtained, including: performing predictive training on the extraction of the position features of the sample set with invalid features removed by using each position predictive model to obtain a position feature set; calculating probability values between each position feature in the position feature set and a preset number of position tags by using a preset activation function, and selecting the position tag with the maximum probability value as a prediction result; judging whether an error value between the predicted result and a real result of the sample set after invalid features are removed meets the preset condition or not by using a preset loss function; if the error value does not meet the preset condition, adjusting the parameter value of each position prediction model, and returning to the step of extracting the position characteristics of the sample set with invalid characteristics removed by using each position prediction model; and if the error value meets the preset condition, exiting the predictive training.
In the embodiment of the invention, the position information in the sample set after the invalid features are removed and the time information corresponding to each position information can be subjected to vector conversion by adopting Glove (Global Vectors for Word Representation, global word vector), embedding Layer and other methods to form a position vector matrix containing the position and time features.
Further, converting the position vector matrix into a feature vector correlation matrix by utilizing a multi-head attention mechanism in each position prediction model; connecting a position vector matrix and the feature vector incidence matrix by using residual error connecting layers in each position prediction model to obtain a position feature vector close incidence matrix; and performing dimension reduction processing on the position feature vector close association matrix by using a full connection layer in each position prediction model to obtain the position feature set.
In the embodiment of the invention, the position features comprise position keyword features, position occurrence time features, same-time same-position occurrence frequency features and the like.
In the embodiment of the present invention, the preset activation functions include, but are not limited to, a softmax activation function, a sigmoid activation function, and a relu activation function.
In the embodiment of the present invention, the preset loss function may be the following function:
Wherein rmse is the error value, num is the number of the sample sets, i is the ith sample in the sample sets, pre i is the predicted result of the ith sample, and grt i is the true result of the ith sample.
S5, calculating a reconciliation value between the precision rate and the recall rate of each position prediction model by using the prediction result and the real result of the sample set, and selecting a position prediction model with the highest reconciliation value as a target position prediction model;
In the embodiment of the present invention, the precision ratio refers to a ratio of the number of samples, where the number of samples corresponds to the predicted result, to the real result, and the predicted result, and the number of samples corresponding to the predicted result, for example, the number of samples, where the position label corresponding to the predicted result is a home address, is 1000, where the position label corresponding to the real result is 300 samples of the home address, and the position labels corresponding to the real results of other 700 samples are not home addresses, and the corresponding precision ratio is 300/1000=30%.
The recall rate refers to a ratio of the number of samples with a predicted result being a to the number of samples with a real result being a for the sample set, for example, the number of samples with a position label corresponding to a preset result being a working address being 1000 and the number of samples with a position label corresponding to a real result being a working address being 20000, and the recall rate is 1000/20000=5%.
In the embodiment of the invention, the following harmonic value calculation formula is adopted to calculate the harmonic value between the precision rate and the recall rate:
In detail, the calculating, by using the prediction result and the real result of the sample set, a harmonic value between the precision rate and the recall rate of each position prediction model includes: obtaining a predicted result and a real result of each sample in the sample set; randomly selecting a prediction result as a reference position; in each position prediction model, respectively counting a first sample number of which the real result and the predicted result in the sample set are both the reference position, a second sample number of which the real result in the sample set of which the predicted result is the reference position is not the reference position, and a third sample number of which the predicted result in the sample set of which the real result is the reference position is not the reference position; calculating the precision rate of each position prediction model according to a precision rate calculation formula by using the first sample number and the second sample number; calculating the recall rate of each position prediction model according to a recall rate calculation formula by using the first sample number and the third sample number; and calculating the reconciliation value between the precision rate and the recall rate of each position prediction model according to a reconciliation value calculation formula by utilizing the precision rate and the recall rate.
In the embodiment of the invention, the calculation formula of the precision ratio is as follows:
in the embodiment of the invention, the recall ratio calculation formula is as follows:
in the embodiment of the invention, the position prediction model with the highest harmonic value is selected as the target position prediction model by calculating the harmonic value of each position prediction model, so that the accuracy of position prediction can be improved.
S6, acquiring historical activity information of the user to be predicted, and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model.
In the embodiment of the invention, the historical activity information of the user to be predicted comprises the position information of the user to be predicted within a certain time period and the time information of the corresponding position information.
In the embodiment of the invention, the target position prediction model has both precision and recall, the position characteristics of the historical activity information are extracted through the target position prediction model, and the position label corresponding to the position characteristics is calculated.
According to the embodiment of the invention, through obtaining the sample set containing the historical activity information of the user position information and the time information, carrying out normalization processing on the time information and the position information in the sample set, removing invalid features in the sample set, improving the quality of the sample set, simultaneously carrying out position prediction training on a preset number of position prediction models by utilizing the sample set, calculating to obtain a reconciliation value between the precision rate and the recall rate of each position prediction model, selecting the position prediction model with the highest reconciliation value as a target position prediction model, and carrying out different position predictions on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model, so that the problem that the geographic position information of the user cannot be obtained when the user is in an offline state can be effectively solved.
Fig. 2 is a functional block diagram of an offline position prediction apparatus according to an embodiment of the present invention.
The offline position predicting apparatus 100 of the present invention may be installed in an electronic device. Depending on the functionality implemented, the offline position prediction apparatus 100 may include a sample processing module 101, a model prediction training module 102, a target model selection module 103, and a target model application module 104. The module of the invention, which may also be referred to as a unit, refers to a series of computer program segments, which are stored in the memory of the electronic device, capable of being executed by the processor of the electronic device and of performing a fixed function.
In the present embodiment, the functions concerning the respective modules/units are as follows:
The sample processing module 101 is configured to obtain a sample set of historical activity information of a user, where the historical activity information includes location information and time information corresponding to the location information, normalize the location information and the time information, and perform an operation of removing invalid features from the normalized sample set;
The model prediction training module 102 is configured to perform prediction training on a preset number of position prediction models by using a sample set after invalid features are removed, until the prediction training meets a preset condition, and quit the prediction training to obtain a prediction result of each position prediction model;
The target model selecting module 103 is configured to calculate a reconciliation value between the precision rate and the recall rate of each of the position prediction models by using the prediction result and the real result of the sample set, and select a position prediction model with the highest reconciliation value as a target position prediction model;
the target model application module 104 is configured to obtain historical activity information of a user to be predicted, and perform location prediction on the user to be predicted according to the historical activity information of the user to be predicted by using the target location prediction model.
In detail, each module in the offline position predicting device 100 in the embodiment of the present invention adopts the same technical means as the offline position predicting method described in fig. 1 and can produce the same technical effects when in use, and will not be described herein.
Fig. 3 is a schematic structural diagram of an electronic device for implementing an offline position prediction method according to an embodiment of the present invention.
The electronic device 1 may comprise a processor 10, a memory 11 and a bus, and may further comprise a computer program, such as an offline position prediction program, stored in the memory 11 and executable on the processor 10.
The memory 11 includes at least one type of readable storage medium, including flash memory, a mobile hard disk, a multimedia card, a card memory (e.g., SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. The memory 11 may in some embodiments be an internal storage unit of the electronic device 1, such as a removable hard disk of the electronic device 1. The memory 11 may in other embodiments also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (SMART MEDIA CARD, SMC), a Secure Digital (SD) card, a flash memory card (FLASH CARD) or the like, which are provided on the electronic device 1. Further, the memory 11 may also include both an internal storage unit and an external storage device of the electronic device 1. The memory 11 may be used not only for storing application software installed in the electronic device 1 and various types of data, such as codes of an offline position prediction program, but also for temporarily storing data that has been output or is to be output.
The processor 10 may be comprised of integrated circuits in some embodiments, for example, a single packaged integrated circuit, or may be comprised of multiple integrated circuits packaged with the same or different functions, including one or more central processing units (Central Processing unit, CPU), microprocessors, digital processing chips, graphics processors, combinations of various control chips, and the like. The processor 10 is a Control Unit (Control Unit) of the electronic device, connects various components of the entire electronic device using various interfaces and lines, executes programs or modules (e.g., an offline position prediction program, etc.) stored in the memory 11 by running or executing the programs or modules, and invokes data stored in the memory 11 to perform various functions of the electronic device 1 and process the data.
The bus may be a peripheral component interconnect standard (PERIPHERAL COMPONENT INTERCONNECT, PCI) bus, or an extended industry standard architecture (extended industry standard architecture, EISA) bus, among others. The bus may be classified as an address bus, a data bus, a control bus, etc. The bus is arranged to enable a connection communication between the memory 11 and at least one processor 10 etc.
Fig. 3 shows only an electronic device with components, it being understood by a person skilled in the art that the structure shown in fig. 3 does not constitute a limitation of the electronic device 1, and may comprise fewer or more components than shown, or may combine certain components, or may be arranged in different components.
For example, although not shown, the electronic device 1 may further include a power source (such as a battery) for supplying power to each component, and preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that functions of charge management, discharge management, power consumption management, and the like are implemented through the power management device. The power supply may also include one or more of any of a direct current or alternating current power supply, recharging device, power failure detection circuit, power converter or inverter, power status indicator, etc. The electronic device 1 may further include various sensors, bluetooth modules, wi-Fi modules, etc., which will not be described herein.
Further, the electronic device 1 may also comprise a network interface, optionally the network interface may comprise a wired interface and/or a wireless interface (e.g. WI-FI interface, bluetooth interface, etc.), typically used for establishing a communication connection between the electronic device 1 and other electronic devices.
The electronic device 1 may optionally further comprise a user interface, which may be a Display, an input unit, such as a Keyboard (Keyboard), or a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, or the like. The display may also be referred to as a display screen or display unit, as appropriate, for displaying information processed in the electronic device 1 and for displaying a visual user interface.
It should be understood that the embodiments described are for illustrative purposes only and are not limited to this configuration in the scope of the patent application.
The offline position prediction program stored in the memory 11 of the electronic device 1 is a combination of instructions that, when executed in the processor 10, may implement:
acquiring a sample set of user historical activity information, wherein the historical activity information comprises position information and time information corresponding to the position information;
Respectively carrying out normalization processing on the position information and the time information;
Performing invalid characteristic eliminating operation on the normalized sample set;
Respectively carrying out predictive training on a preset number of position predictive models by using a sample set after invalid features are removed, and exiting the predictive training until the predictive training meets preset conditions to obtain a predictive result of each position predictive model;
Calculating a reconciliation value between the precision rate and the recall rate of each position prediction model by using the prediction result and the real result of the sample set, and selecting a position prediction model with the highest reconciliation value as a target position prediction model;
and acquiring historical activity information of the user to be predicted, and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model.
Specifically, the specific implementation method of the above instructions by the processor 10 may refer to the description of the relevant steps in the corresponding embodiment of fig. 1, which is not repeated herein.
Further, the modules/units integrated in the electronic device 1 may be stored in a computer readable storage medium if implemented in the form of software functional units and sold or used as separate products. The computer readable storage medium may be volatile or nonvolatile. For example, the computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM).
The present invention also provides a computer readable storage medium storing a computer program which, when executed by a processor of an electronic device, can implement:
acquiring a sample set of user historical activity information, wherein the historical activity information comprises position information and time information corresponding to the position information;
Respectively carrying out normalization processing on the position information and the time information;
Performing invalid characteristic eliminating operation on the normalized sample set;
Respectively carrying out predictive training on a preset number of position predictive models by using a sample set after invalid features are removed, and exiting the predictive training until the predictive training meets preset conditions to obtain a predictive result of each position predictive model;
Calculating a reconciliation value between the precision rate and the recall rate of each position prediction model by using the prediction result and the real result of the sample set, and selecting a position prediction model with the highest reconciliation value as a target position prediction model;
and acquiring historical activity information of the user to be predicted, and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model.
In the several embodiments provided in the present invention, it should be understood that the disclosed apparatus, device and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is merely a logical function division, and there may be other manners of division when actually implemented.
The modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional module in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units can be realized in a form of hardware or a form of hardware and a form of software functional modules.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof.
The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
The blockchain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, encryption algorithm and the like. The blockchain (Blockchain), essentially a de-centralized database, is a string of data blocks that are generated in association using cryptographic methods, each of which contains information from a batch of network transactions for verifying the validity (anti-counterfeit) of its information and generating the next block. The blockchain may include a blockchain underlying platform, a platform product services layer, an application services layer, and the like.
The embodiment of the application can acquire and process the related data based on the artificial intelligence technology. Wherein artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) is the theory, method, technique, and application system that uses a digital computer or a digital computer-controlled machine to simulate, extend, and expand human intelligence, sense the environment, acquire knowledge, and use knowledge to obtain optimal results.
Furthermore, it is evident that the word "comprising" does not exclude other elements or steps, and that the singular does not exclude a plurality. A plurality of units or means recited in the system claims can also be implemented by means of software or hardware by means of one unit or means. The terms second, etc. are used to denote a name, but not any particular order.
Finally, it should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.

Claims (8)

1. An offline position prediction method, the method comprising:
acquiring a sample set of user historical activity information, wherein the historical activity information comprises position information and time information corresponding to the position information;
Respectively carrying out normalization processing on the position information and the time information;
Performing invalid characteristic eliminating operation on the normalized sample set;
Respectively carrying out predictive training on a preset number of position predictive models by using a sample set after invalid features are removed, and exiting the predictive training until the predictive training meets preset conditions to obtain a predictive result of each position predictive model;
Calculating a reconciliation value between the precision rate and the recall rate of each position prediction model by using the prediction result and the real result of the sample set, and selecting a position prediction model with the highest reconciliation value as a target position prediction model;
Acquiring historical activity information of a user to be predicted, and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model;
The method for predicting the position prediction models by using the sample set after invalid features are removed includes the steps of: performing predictive training on the extraction of the position features of the sample set with invalid features removed by using each position predictive model to obtain a position feature set; calculating probability values between each position feature in the position feature set and a preset number of position tags by using a preset activation function, and selecting the position tag with the maximum probability value as a prediction result; judging whether an error value between the predicted result and a real result of the sample set after invalid features are removed meets the preset condition or not by using a preset loss function; if the error value does not meet the preset condition, adjusting the parameter value of each position prediction model, and returning to the step of extracting the position characteristics of the sample set with invalid characteristics removed by using each position prediction model; if the error value meets the preset condition, exiting the predictive training;
The calculating, by using the prediction result and the real result of the sample set, a harmonic value between the precision rate and the recall rate of each position prediction model includes: obtaining a predicted result and a real result of each sample in the sample set; randomly selecting a prediction result as a reference position; in each position prediction model, respectively counting a first sample number of which the real result and the predicted result in the sample set are both the reference position, a second sample number of which the real result in the sample set of which the predicted result is the reference position is not the reference position, and a third sample number of which the predicted result in the sample set of which the real result is the reference position is not the reference position; calculating the precision rate of each position prediction model according to a precision rate calculation formula by using the first sample number and the second sample number; calculating the recall rate of each position prediction model according to a recall rate calculation formula by using the first sample number and the third sample number; and calculating the reconciliation value between the precision rate and the recall rate of each position prediction model according to a reconciliation value calculation formula by utilizing the precision rate and the recall rate.
2. The offline position prediction method according to claim 1, wherein the normalizing the position information and the time information respectively includes:
Marking the time information in the sample set by using a preset time tag to obtain marked time information;
calculating the similarity between the position information, and determining the position information corresponding to the similarity larger than a preset similarity threshold value as the same position set;
And randomly selecting one position information from the same position set to replace each position information in the same position set.
3. The offline position prediction method of claim 2, wherein said calculating a similarity between each of said position information comprises:
identifying longitude and latitude corresponding to each position information in the sample set;
generating GeoHash codes of each piece of position information by utilizing GeoHash algorithm according to the longitude and latitude of each piece of position information;
And calculating the coding similarity between GeoHash codes of each piece of position information, and taking the coding similarity as the similarity between the corresponding pieces of position information.
4. The offline position prediction method according to claim 1, wherein the performing an invalid feature rejection operation on the normalized sample set includes:
Extracting the position characteristics of the corresponding samples of each user in the sample set one by one;
Sequentially calculating the percentage between the sample number corresponding to each position feature of each user and the total sample number corresponding to the user to obtain the coverage rate of each position feature of the user;
And determining the position features with the coverage rate smaller than a preset coverage rate threshold as invalid features and eliminating the invalid features.
5. The offline position prediction method of claim 1, wherein the step of performing an invalid feature rejection operation on the normalized sample set is replaced by performing an invalid sample rejection operation on the normalized sample set, the invalid sample rejection operation on the normalized sample set comprising:
respectively pre-training a preset number of position prediction models by using the sample set, and exiting the pre-training until the pre-training meets the preset training times to obtain a pre-training result of each position prediction model;
counting the repetition rate of the pre-training result of each sample in the sample set in each position prediction model;
and taking the sample with the repetition rate smaller than a preset repetition rate threshold as an invalid sample, and removing the invalid sample from the sample set.
6. An offline position prediction apparatus for implementing the offline position prediction method according to any one of claims 1 to 5, characterized in that the apparatus comprises:
the sample processing module is used for acquiring a sample set of user historical activity information, wherein the historical activity information comprises position information and time information corresponding to the position information, normalization processing is carried out on the position information and the time information respectively, and invalid characteristic eliminating operation is carried out on the normalized sample set;
The model prediction training module is used for respectively carrying out prediction training on a preset number of position prediction models by utilizing a sample set after invalid features are removed, and stopping the prediction training until the prediction training meets preset conditions to obtain a prediction result of each position prediction model;
The target model selection module is used for calculating a reconciliation value between the precision rate and the recall rate of each position prediction model by utilizing the prediction result and the real result of the sample set, and selecting the position prediction model with the highest reconciliation value as a target position prediction model;
And the target model application module is used for acquiring the historical activity information of the user to be predicted, and carrying out position prediction on the user to be predicted according to the historical activity information of the user to be predicted by utilizing the target position prediction model.
7. An electronic device, the electronic device comprising:
At least one processor; and
A memory communicatively coupled to the at least one processor; wherein,
The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the offline position prediction method of any one of claims 1 to 5.
8. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the off-line position prediction method according to any one of claims 1 to 5.
CN202210583200.9A 2022-05-25 2022-05-25 Offline position prediction method and device, electronic equipment and storage medium Active CN114925920B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210583200.9A CN114925920B (en) 2022-05-25 2022-05-25 Offline position prediction method and device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210583200.9A CN114925920B (en) 2022-05-25 2022-05-25 Offline position prediction method and device, electronic equipment and storage medium

Publications (2)

Publication Number Publication Date
CN114925920A CN114925920A (en) 2022-08-19
CN114925920B true CN114925920B (en) 2024-05-03

Family

ID=82809719

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210583200.9A Active CN114925920B (en) 2022-05-25 2022-05-25 Offline position prediction method and device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN114925920B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103246687A (en) * 2012-06-13 2013-08-14 苏州大学 Automatic Blog abstracting method based on characteristic information
WO2020133960A1 (en) * 2018-12-25 2020-07-02 平安科技（深圳）有限公司 Text quality inspection method, electronic apparatus, computer device and storage medium
CN111882339A (en) * 2019-12-20 2020-11-03 马上消费金融股份有限公司 Prediction model training and response rate prediction method, device, equipment and storage medium
WO2021051596A1 (en) * 2019-09-19 2021-03-25 平安科技（深圳）有限公司 Model prediction optimization method, apparatus and device, and readable storage medium
WO2021259007A1 (en) * 2020-06-24 2021-12-30 平安科技（深圳）有限公司 Data pushing method and system based on behavior trajectory, and computer device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103246687A (en) * 2012-06-13 2013-08-14 苏州大学 Automatic Blog abstracting method based on characteristic information
WO2020133960A1 (en) * 2018-12-25 2020-07-02 平安科技（深圳）有限公司 Text quality inspection method, electronic apparatus, computer device and storage medium
WO2021051596A1 (en) * 2019-09-19 2021-03-25 平安科技（深圳）有限公司 Model prediction optimization method, apparatus and device, and readable storage medium
CN111882339A (en) * 2019-12-20 2020-11-03 马上消费金融股份有限公司 Prediction model training and response rate prediction method, device, equipment and storage medium
WO2021259007A1 (en) * 2020-06-24 2021-12-30 平安科技（深圳）有限公司 Data pushing method and system based on behavior trajectory, and computer device

Also Published As

Publication number Publication date
CN114925920A (en) 2022-08-19

Similar Documents

Publication Publication Date Title
CN111506723B (en) Question-answer response method, device, equipment and storage medium
US9852177B1 (en) System and method for generating automated response to an input query received from a user in a human-machine interaction environment
CN113807973B (en) Text error correction method, apparatus, electronic device and computer readable storage medium
CN113626606B (en) Information classification method, device, electronic equipment and readable storage medium
CN112269875B (en) Text classification method, device, electronic equipment and storage medium
CN111738778B (en) User portrait generation method and device, computer equipment and storage medium
CN114880449B (en) Method and device for generating answers of intelligent questions and answers, electronic equipment and storage medium
CN114399212A (en) Ecological environment quality evaluation method and device, electronic equipment and storage medium
CN113807728A (en) Performance assessment method, device, equipment and storage medium based on neural network
CN114595321A (en) Question marking method and device, electronic equipment and storage medium
CN114022841A (en) Personnel monitoring and identifying method and device, electronic equipment and readable storage medium
CN116401606A (en) Fraud identification method, device, equipment and medium
CN112651782A (en) Behavior prediction method, device, equipment and medium based on zoom dot product attention
CN114925920B (en) Offline position prediction method and device, electronic equipment and storage medium
CN116843395A (en) Alarm classification method, device, equipment and storage medium of service system
CN113806540B (en) Text labeling method, text labeling device, electronic equipment and storage medium
CN116737947A (en) Entity relationship diagram construction method, device, equipment and storage medium
CN113706019B (en) Service capability analysis method, device, equipment and medium based on multidimensional data
CN113919962B (en) Intelligent analysis method, device, equipment and storage medium for insurance policy
CN113902418B (en) Project approval method and device, electronic equipment and readable storage medium
CN113992745B (en) Activity information pushing method and device, electronic equipment and storage medium
CN113792342B (en) Desensitization data reduction method, device, computer equipment and storage medium
CN114546882B (en) Intelligent question-answering system testing method and device, electronic equipment and storage medium
CN113825138B (en) Fraud short message monitoring method and device, electronic equipment and storage medium
CN113673597B (en) Enterprise annual report prompting method, device, electronic equipment and computer 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
GR01 Patent grant
GR01 Patent grant