Gas station information real-time display method and device and terminal equipment

Abstract

The embodiment of the invention provides a gas station information real-time display method and device and terminal equipment, and belongs to the technical field of data processing. The method comprises the following steps: clustering historical evaluation information and target information to obtain a clustering result, and determining a first recommendation parameter according to the clustering result; obtaining related information of the surrounding objects, and determining a second recommendation parameter according to the related information and the target information; acquiring historical vehicle flow of an associated road section between the target object and the surrounding object according to a sensor; obtaining an association relationship between the associated road sections according to the historical vehicle flow; determining a flow prediction model according to the association relationship among the historical vehicle flow, the historical user flow and the associated road sections in combination with the first recommendation parameter and the second recommendation parameter; performing user traffic prediction on the target object according to the traffic prediction model to obtain a target prediction result; and determining a target state of the target object according to the target prediction result, and performing real-time display according to the target state to obtain a target display result.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device and terminal equipment for real-time display of gas station information. Background Art

[0002] In the gas station operation scenario, real-time large-screen display has become a common and important way to present information. However, at present, most gas stations have a relatively simple function when using real-time large screens to display data, which is just a simple listing and display of existing or directly accessible data. This method is limited to direct data display, so that the real-time large screen fails to fully realize its potential value. The real-time display method of gas stations in the existing technology cannot provide more in-depth and forward-looking decision-making basis for gas station managers, which is not conducive to improving gas stations' operational efficiency and service quality. Summary of the invention

[0003] The main purpose of the embodiments of the present invention is to provide a method, device and terminal device for real-time display of gas station information, aiming to solve the problem that the real-time display method of the related art cannot provide more in-depth and forward-looking decision-making basis for gas station managers, which is not conducive to improving gas stations' operational efficiency and service quality.

[0004] In a first aspect, an embodiment of the present invention provides a method for real-time display of gas station information, comprising:

[0005] Obtain the historical user traffic and historical evaluation information of the target object, and obtain the target information corresponding to the target object under social media;

[0006] Clustering the historical evaluation information and the target information to obtain a clustering result, and determining a first recommendation parameter according to the clustering result;

[0007] Obtaining relevant information corresponding to surrounding objects of the target object, and determining a second recommendation parameter according to the relevant information and the target information;

[0008] Determine an associated road section according to the target object and the surrounding objects, and obtain a historical vehicle flow of the associated road section according to a sensor;

[0009] Obtaining corresponding association relationships between the associated road sections according to the historical vehicle flow;

[0010] Determine a traffic prediction model according to the historical vehicle traffic, the historical user traffic, and the corresponding association relationship between the associated road sections in combination with the first recommended parameter and the second recommended parameter;

[0011] Performing user traffic prediction on the target object according to the traffic prediction model to obtain a target prediction result;

[0012] The target state of the target object is determined according to the target prediction result, and a target display result is obtained by performing real-time display according to the target state.

[0013] In a second aspect, an embodiment of the present invention provides a gas station information real-time display device, comprising:

[0014] The data acquisition module is used to obtain the historical user traffic and historical evaluation information of the target object, and obtain the target information corresponding to the target object under social media;

[0015] A data clustering module, used for clustering the historical evaluation information and the target information to obtain a clustering result, and determining a first recommendation parameter according to the clustering result;

[0016] A parameter determination module, used to obtain relevant information corresponding to the surrounding objects of the target object, and determine a second recommended parameter according to the relevant information and the target information;

[0017] A traffic determination module, used to determine an associated road section according to the target object and the surrounding objects, and obtain a historical vehicle traffic of the associated road section according to a sensor;

[0018] A relationship determination module, used for obtaining the corresponding association relationship between the associated road sections according to the historical vehicle flow;

[0019] A model determination module, configured to determine a traffic prediction model according to the historical vehicle traffic, the historical user traffic and the corresponding association relationship between the associated road sections in combination with the first recommended parameter and the second recommended parameter;

[0020] A data prediction module, used to perform user traffic prediction on the target object according to the traffic prediction model to obtain a target prediction result;

[0021] The data display module is used to determine the target state of the target object according to the target prediction result, and to display the target state in real time to obtain a target display result.

[0022] In a third aspect, an embodiment of the present invention further provides a terminal device, comprising a processor, a memory, a computer program stored in the memory and executable by the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein when the computer program is executed by the processor, the steps of any one of the methods for real-time display of gas station information provided in the specification of the present invention are implemented.

[0023] The embodiment of the present invention provides a method, device and terminal device for real-time display of gas station information, the method comprising: obtaining historical user traffic and historical evaluation information of a target object, and obtaining target information corresponding to the target object under social media, so as to understand the internal operating conditions of the target object and its external environment, so as to provide support for providing more accurate prediction results in the future, clustering the historical evaluation information and target information to obtain clustering results, and determining a first recommended parameter according to the clustering results; thereby obtaining relevant information corresponding to surrounding objects of the target object, and determining a second recommended parameter according to the relevant information and the target information, so as to identify the potential impact of external factors on the traffic of the target object; thereby determining associated sections according to the target object and the surrounding objects, and obtaining the historical vehicle traffic of the associated sections according to sensors, and obtaining the corresponding correlation relationship between the associated sections according to the historical vehicle traffic; thereby determining a traffic prediction model according to the historical vehicle traffic, the historical user traffic and the corresponding correlation relationship between the associated sections in combination with the first recommended parameter and the second recommended parameter, so that the model fully considers factors in multiple aspects and can more accurately simulate the user traffic changes of the target object. By learning and training a large amount of historical data, the model can capture the complex relationship between various factors, thereby providing users with accurate prediction results, and then predicting the user traffic of the target object according to the traffic prediction model to obtain the target prediction result. Finally, the target state of the target object is determined according to the target prediction result, and the target state is displayed in real time to obtain the target display result, so that the gas station manager can make resource allocation in advance according to the target display result and improve the operation efficiency of the gas station. At the same time, it can better meet the needs of users, improve service quality, and enhance user satisfaction. It also solves the problem that the real-time display method of the relevant technology cannot provide more in-depth and forward-looking decision-making basis for gas station managers, which is not conducive to improving the operation efficiency and service quality of gas stations, and improves the operation efficiency and service quality of target gas stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 A flow chart of a method for real-time display of gas station information provided by an embodiment of the present invention;

[0026] Figure 2 A schematic diagram of the module structure of a gas station information real-time display device provided by an embodiment of the present invention;

[0027] Figure 3A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0030] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0031] The embodiments of the present invention provide a method, device and terminal device for real-time display of gas station information. The method for real-time display of gas station information can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant and a wearable device. The terminal device can be a server or a server cluster.

[0032] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0033] Please refer to Figure 1 , Figure 1 A flowchart of a method for real-time display of gas station information provided by an embodiment of the present invention.

[0034] like Figure 1 As shown, the method for real-time display of gas station information includes steps S101 to S108.

[0035] Step S101: Obtain historical user traffic and historical evaluation information of a target object, and obtain target information corresponding to the target object in social media.

[0036] Exemplarily, the target object is a gas station to be analyzed. First, the user flow data corresponding to the target object is collected from the management system corresponding to the target object and stored in the corresponding database. Then, the historical user flow of the target object is obtained according to the database query statement, where the historical user flow may refer to the number of vehicles entering the target object per day in a historical time period, and the historical evaluation information of users on the target object is extracted from the customer feedback platform.

[0037] For example, keywords related to gas stations are used to search on public social media such as TikTok, Weibo, or local life service platforms or community forums, such as the name, address, brand, etc. of the gas station, to obtain the target information corresponding to the evaluation of the target object by users on social media.

[0038] In some embodiments, obtaining target information corresponding to a target object on social media includes: obtaining a first position corresponding to the target object, and collecting first information corresponding to the target object on social media based on the first position, wherein the first information is used to characterize user-published information corresponding to the social media and the first position, and the first information includes first text information and first image information; performing target association from the first text information to obtain relevant text information associated with the target object, and performing target classification on the first image information to obtain an image description object corresponding to the first image information; obtaining relevant image information associated with the target object from the first image information based on the image description object; performing position prediction on the first information based on the relevant text information and the relevant image information to obtain relevant position information corresponding to the first information; and screening the first information based on the relevant position information and the first position to obtain the target information corresponding to the target object.

[0039] Exemplarily, the longitude and latitude information corresponding to the target object is directly acquired as the first position through a geographic positioning system or map data.

[0040] Exemplarily, data is collected on the social media platform based on the first location and the target object using an open interface or data capture tool of the social media (under the premise of legality and compliance). For example, for social media with a location tag function, the search range is set to a certain radius area centered on the first location, and the first information related to the target object is searched, wherein the first information is used to characterize the information posted by the user corresponding to the social media and the first location, and the first information includes first text information and first image information.

[0041] Exemplarily, the content related to the target object in the first text information is extracted using methods such as keyword matching and semantic understanding to obtain relevant text information associated with the target object. The first image information is classified using computer vision technology. The subject, scene, object and other elements in the image are identified through a trained image classification model, thereby determining the object described by the image.

[0042] Exemplarily, the first image information is further filtered according to the image description object. If the image description object is associated with the target object, the image in the image description object that is associated with the target object is used as the relevant image information associated with the target object.

[0043] Exemplarily, the location prediction is performed in combination with the relevant text information and the relevant image information. For the relevant text information, the place name, direction description, distance information, etc. mentioned therein are analyzed; for the relevant image information, the location can be inferred through the geographical features in the image (such as landmark buildings, road signs, etc.) or the shooting parameters of the image (such as GPS tags), so as to use the geographic information system technology to combine these information to perform location prediction on the first information and obtain the relevant location information corresponding to the first information.

[0044] Exemplarily, the first information is screened according to the relevant position information and the first position. If the relevant position information and the first position are within a certain error range (the error range can be set according to actual needs), the information is considered to be closely related to the target object and is used as the target information corresponding to the target object.

[0045] Specifically, by processing and screening the first information, information related to the target object can be accurately obtained from massive social media data. Interference from irrelevant information is avoided, and the quality and pertinence of the information are improved. By predicting and screening the location, the location correlation between the obtained target information and the target object is ensured.

[0046] In some embodiments, the performing position prediction on the first information according to the relevant text information and the relevant image information to obtain the relevant position information corresponding to the first information includes: performing image feature extraction on the relevant image information using the image coding layer of the position prediction model to obtain a first feature vector; performing text feature extraction on the relevant text information using the text coding layer of the position prediction model to obtain a second feature vector; determining the information matching value between the relevant image information and the relevant text information according to the first feature vector and the second feature vector using the image-text matching layer of the position prediction model; performing position sign extraction on the relevant image information using the image position determination layer of the position prediction model to obtain the image position information corresponding to the relevant image information; performing position sign extraction on the relevant text information using the text position determination layer of the position prediction model to obtain the text position information corresponding to the relevant text information; performing fusion of the image position information and the text position information using the feature fusion layer of the position prediction model according to the information matching value to obtain a target fusion feature; performing position prediction on the first information using the target fusion feature using the position classification layer of the position prediction model to obtain the relevant position information corresponding to the first information.

[0047] Exemplarily, a position prediction model is constructed, wherein the position prediction model includes an image coding layer, a text coding layer, an image-text matching layer, an image position determination layer, a text position determination layer, a feature fusion layer, and a position classification layer. The image coding layer may use a pre-trained convolutional neural network (CNN), such as ResNet, VGG, etc. Thus, the relevant image information is input into the image coding layer, and these pre-trained CNNs will perform multi-scale feature extraction on the image to obtain a first feature vector, which represents the feature representation of the relevant image information.

[0048] For example, the text encoding layer can use pre-trained language models such as BERT, GPT, etc. Inputting relevant text information into the text encoding layer language model will perform semantic understanding and feature extraction on the text. They will consider the vocabulary, grammar, context and other information in the text and convert the text into a vector representation, namely the second feature vector. The second feature vector can reflect the semantic features of the relevant text information.

[0049] Exemplarily, the image-text matching layer receives the first feature vector and the second feature vector, and then uses a measurement method such as cosine similarity and Euclidean distance to calculate the similarity between the two vectors. The similarity value is used as the information matching value between the relevant image information and the relevant text information. For example, the closer the cosine similarity is to 1, the higher the correlation between the image and the text; the closer it is to 0, the lower the correlation.

[0050] For example, the image location determination layer further analyzes the relevant image information and uses the geographical features in the image, such as landmark buildings, road signs, etc., and the image shooting parameters (such as GPS tags) to extract the image location characteristics. By analyzing and processing these location characteristics, the image location information corresponding to the relevant image information is determined, such as latitude and longitude coordinates or specific address descriptions.

[0051] Exemplarily, the text location determination layer performs natural language processing on the relevant text information. By identifying location-related words and sentences such as place names, direction descriptions, and distance information mentioned in the text, the text location information corresponding to the relevant text information is extracted. For example, if the text mentions "next to the library in the city center", then this location information can be extracted.

[0052] For example, the feature fusion layer will fuse the image position information and the text position information according to the information matching value calculated previously. If the information matching value is high, it means that the image and text are highly correlated, so a higher weight will be given to both during fusion; if the information matching value is low, it means that the correlation is weak, and the weight may be adjusted according to the specific situation to obtain the target fusion feature.

[0053] Exemplarily, the location classification layer receives the target fusion feature and uses the target fusion feature to predict the location of the first information. Classification algorithms such as decision trees, support vector machines, etc., or deep learning models such as fully connected neural networks, etc. can be used. Therefore, according to the patterns and rules learned by the target fusion feature, the location of the first information is predicted, and finally the relevant location information corresponding to the first information is output, such as specific geographical location coordinates or addresses. The relevant location information can also be a judgment result of whether it is located near the target gas station. Therefore, when the relevant location information is a description of being located near the target gas station, the published information corresponding to the relevant location information can be determined as a comment on the target gas station.

[0054] Specifically, by comprehensively utilizing the information of images and text, the location can be analyzed and judged from different angles. Images can provide intuitive geographical features, and text can supplement detailed location descriptions. The combination of the two can provide a more comprehensive understanding of the location characteristics of the information, thereby improving the accuracy of location prediction.

[0055] Step S102: cluster the historical evaluation information and the target information to obtain a clustering result, and determine a first recommendation parameter according to the clustering result.

[0056] Exemplarily, the historical evaluation information is the evaluation content collected by users on the customer feedback platform corresponding to the target object, and the target information is the content of the user's evaluation of the target object on social media, so that the historical evaluation information and the target information are clustered according to clustering algorithms such as DBSCAN and K-means to obtain clustering results.

[0057] Exemplarily, before performing data analysis on each cluster in the clustering results, we must first ensure that the data in the cluster is in a state suitable for analysis, and then perform preprocessing operations on the text data in the cluster, such as removing stop words, performing stem extraction and part-of-speech tagging, so as to use the word frequency inverse document frequency method to find out the words that appear frequently in the cluster and the words with important weights, and then determine the theme of the cluster based on the words that appear frequently and the words with important weights. For example, if words such as "oil price discount" and "gas card activities" appear frequently in a certain cluster, then it can be determined that the theme of the cluster is related to the preferential activities of the gas station; if words such as "good service attitude" and "clean environment" appear frequently, then the theme of the cluster may be related to the service and environment of the gas station.

[0058] For example, after clarifying the theme of each cluster, it is necessary to analyze the relationship between each theme and the user's choice of gas station. Different themes may have different degrees of influence on the user's choice of gas station. For example, for price-sensitive users, the cluster with the theme of "oil price discount" may have a greater impact on their choice; while for users who focus on service experience, the cluster with the theme of "good service attitude" is more important. In addition to the theme of the cluster, the amount of data in the cluster is also an important consideration. The more data in the cluster, the more common the features represented by the cluster are in the overall data, and the greater the impact on the user's choice. Each cluster can be assigned a corresponding weight according to the amount of data. Clusters with a large amount of data have a higher weight, and clusters with a small amount of data have a lower weight.

[0059] Exemplarily, the first recommendation parameter corresponding to the target object is determined by comprehensively considering the themes of each cluster and the corresponding weights. A mathematical model or rule can be established to quantify the theme information and weights of the clusters and obtain a comprehensive value as the first recommendation parameter. For example, a weighted summation method can be used to multiply the influence of each cluster theme on the user's choice by the weight of the cluster, and then add the results of all clusters to obtain the final first recommendation parameter. The first recommendation parameter can be used to characterize the possibility of users choosing the target object among many gas stations. The higher the parameter value, the more the performance of the target object in various aspects meets the needs and preferences of the user, and the greater the possibility of the user choosing the gas station. At the same time, the first recommendation parameter can also be used to characterize the attractiveness of the target object to the user. It combines the information reflected by each cluster and can fully reflect the advantages and characteristics of the target object. A gas station with a higher first recommendation parameter is more attractive in market competition and can attract more users to refuel.

[0060] In some embodiments, clustering the historical evaluation information and the target information to obtain a clustering result includes: text merging the historical evaluation information and the target information to obtain a text set, and performing word segmentation processing on each sub-text information in the text set to obtain associated words corresponding to the sub-text information; determining a first text similarity corresponding to any two of the sub-text information based on the associated words; determining a target dimension, and obtaining a word representation vector corresponding to the associated words, and determining a word dimension association value corresponding to the associated words under the target dimension based on the word representation vector; determining weight information corresponding to the sub-text information under the target dimension based on the word dimension association value; determining a text dimension association value corresponding to the sub-text information when it belongs to the target dimension based on the word dimension association value and the weight information; determining a second text similarity corresponding to any two of the sub-text information based on the text dimension association value; clustering the text set based on the first text similarity and the second text similarity to obtain the clustering result.

[0061] For example, the historical evaluation information and the target information are integrated into a unified text set. This step can centralize the scattered information and more comprehensively reflect various information related to the gas station.

[0062] For example, a word segmentation operation is performed on each sub-text information in the text set. Word segmentation is to split the continuous text into meaningful words, so that the text information can be converted into basic units that are easier to analyze, thereby obtaining the associated words corresponding to each sub-text information.

[0063] Exemplarily, associated words corresponding to any two sub-text information in a text set are obtained, and then the proportion of the number of identical associated words in the two sub-texts to the total number of associated words is counted based on the associated words corresponding to the two sub-text information. The higher the proportion, the higher the similarity, and thus the first text similarity corresponding to the sub-text information is obtained.

[0064] For example, the target dimension is determined, that is, the entire data semantic space where all sub-text information in the text collection is distributed. The target dimension can be one semantic dimension or multiple semantic dimensions, and then the pre-trained word vector model, such as Word2Vec or GloVe, is used to convert the associated words into vector form, that is, the word representation vector. Then, the associated words are mapped to the target dimension by vector decomposition based on the word representation vector, so as to obtain the dimensional representation vector corresponding to the associated words under the target dimension, and then the word dimension association value corresponding to the associated words under the target dimension is determined based on the dimensional representation vector.

[0065] For example, the dimensional representation vector corresponding to each associated word under each target dimension is obtained, and then the dimensional semantic center calculation is performed on all dimensional representation vectors under the target dimension to obtain the dimensional center vector corresponding to the target dimension, and then the distance information between the dimensional representation vector and the dimensional center vector is calculated, and then the word dimension association value corresponding to the associated word under the target dimension is determined according to the distance information.

[0066] Exemplarily, a target threshold is determined, and then the word dimension association values ​​corresponding to the associated words under the target dimension are compared with the target threshold one by one. The word dimension association value reflects the degree of association between the associated words and the target dimension. When the word dimension association value of a certain associated word is greater than the target threshold, it indicates that the associated word has a strong association with the target dimension, and then the associated word can be determined to be a related word of the sub-text information under the target dimension.

[0067] Exemplarily, after determining the relevant words of the sub-text information under the target dimension, the weight information corresponding to the sub-text information under the target dimension is further determined according to the number of these relevant words. The number of relevant words reflects the strength of association between the sub-text information and the target dimension to a certain extent. Generally speaking, the more relevant words there are, the more obvious the characteristics of the sub-text information on the target dimension are, and the higher the corresponding weight is.

[0068] Exemplarily, the text dimension association value when the sub-text information belongs to the target dimension is calculated by weighted summation based on the weight information and the word dimension association value. Specifically, the word dimension association value of each related word is multiplied by its corresponding weight, and then all the results are added together. The sum obtained is the text dimension association value of the sub-text information under the target dimension. This value comprehensively considers the degree of association of the related words and the importance of the sub-text information under the target dimension, and can more comprehensively and accurately reflect the association between the sub-text information and the target dimension.

[0069] Exemplarily, after obtaining the text dimension association value corresponding to each sub-text information under each target dimension, the text dimension association values ​​under all target dimensions can be further summed. By adding the text dimension association values ​​of each sub-text information under each target dimension, the text information representation value corresponding to the sub-text information can be obtained. This representation value is a comprehensive reflection of the sub-text information under multiple target dimensions, which can more comprehensively reflect the characteristics and connotations of the sub-text information.

[0070] Exemplarily, on this basis, the difference between the text information representation values ​​of any two sub-text information is calculated, and this difference is the second text similarity between any two sub-text information. Different from the traditional text similarity calculation method, this similarity fully considers the degree of association between the texts under the target dimension. It not only focuses on the word overlap on the surface of the text, but also digs deeper into the intrinsic connection between the texts in different target dimensions. Therefore, the second text similarity can more deeply and accurately reflect the similarity between texts, and provide a more reliable basis for subsequent tasks such as text clustering, information retrieval, and data analysis.

[0071] Exemplarily, a weighted average or other method is used to combine the first text similarity and the second text similarity to obtain a comprehensive similarity index, and then a clustering algorithm such as hierarchical clustering is used to divide the text collection into different clusters based on the comprehensive similarity index. The texts in each cluster have a high similarity, thereby obtaining a clustering result.

[0072] Specifically, the first text similarity and the second text similarity are calculated and clustered by combining the two, taking into account multiple aspects of the text. The first text similarity measures text similarity from the perspective of word overlap, while the second text similarity combines the information of the target dimension to more deeply explore the association of texts under specific dimensions. This multi-dimensional consideration makes the clustering results more accurate and comprehensive, and can better reflect the true relationship between texts.

[0073] In some embodiments, the determining of the first text similarity between any two of the subtext information according to the associated words includes: obtaining any two of the subtext information as a first subtext and a second subtext, and obtaining a first word corresponding to the first subtext and a second word corresponding to the second subtext; determining a first part of speech and a first word position corresponding to the first word in the first subtext, and determining a first weight corresponding to the first word according to the first part of speech and the first word position; determining a second part of speech and a second word position corresponding to the second word in the second subtext, and determining a second weight corresponding to the second word according to the second part of speech and the second word position; determining a first key degree corresponding to the first word according to a first frequency corresponding to the first word in the text set; determining a second key degree corresponding to the second word according to a second frequency corresponding to the second word in the text set; performing text similarity calculation according to the first word and the second word to obtain word similarity, and fusing the word similarity according to the first weight, the second weight, the first key degree and the second key degree to obtain the first similarity between the first subtext and the second subtext; performing data adjustment on the first similarity according to the word similarity to obtain the first text similarity between the first subtext and the second subtext; wherein the first similarity is obtained according to the following formula:

[0074]

[0075] in, represents the first similarity between the i-th first subtext and the j-th second subtext, represents the number of the first words corresponding to the i-th first subtext; represents the number of the second words corresponding to the j-th second subtext; represents the word similarity between the tth first word in the i-th first subtext and the hth second word in the j-th second subtext, represents the first key degree corresponding to the t-th first word in the i-th first subtext, represents the first weight corresponding to the t-th first word in the i-th first subtext, represents the second key degree corresponding to the hth second word of the jth second subtext, represents the second weight corresponding to the hth second word of the jth second subtext, represents the similarity between the mth first word and the nth first word in the i-th first subtext, represents the first key degree corresponding to the mth first word in the i-th first subtext, represents the first weight corresponding to the mth first word of the i-th first subtext, represents the first key degree corresponding to the nth first word in the i-th first subtext, represents the first weight corresponding to the nth first word in the i-th first subtext, represents the similarity between the kth second word and the qth second word of the jth second subtext, represents the second key degree corresponding to the kth second word of the jth second subtext, represents the second weight corresponding to the kth second word of the jth second subtext, The second key degree corresponding to the qth second word of the jth second subtext, Indicates the second weight corresponding to the qth second word in the jth second subtext.

[0076] Exemplarily, any two sub-text information are selected from the text set as the first sub-text and the second sub-text, respectively. Then, the two sub-texts are segmented to obtain a first word corresponding to the first sub-text and a second word corresponding to the second sub-text.

[0077] Exemplarily, for the first word, its part of speech (such as noun, verb, adjective, etc.) and position in the first subtext are determined. Different parts of speech have different importance in text expression. For example, nouns are usually the core objects of text description and are relatively important; while some function words may be less important. The position of a word will also affect its weight. For example, words at the beginning or end of a sentence may receive more attention and their weight will increase accordingly. Considering the first part of speech and the first word position, determine the corresponding first weight for the first word. The same method is applied to the second word in the second subtext. Determine the second part of speech and the second word position of the second word, and then determine the second weight corresponding to the second word based on this information.

[0078] Exemplarily, the frequency of occurrence of the first word in the entire text collection is counted, that is, the first frequency. Words with a higher frequency of occurrence may be some common words with relatively low criticality; while words with a lower frequency of occurrence but playing a key role in a specific subtext have a higher criticality. The first criticality corresponding to the first word is determined based on the first frequency. In the same way, the second frequency of the second word in the text collection is counted, and the second criticality corresponding to the second word is determined accordingly.

[0079] Exemplarily, a pre-trained word vector model is used to perform word vector representation on the first word and the second word to obtain a first vector corresponding to the first word and a second vector corresponding to the second word, and then the semantic similarity between the first word and the second word is determined based on the first vector and the second vector using the cosine formula.

[0080] Exemplarily, the first similarity between the first subtext and the second subtext is obtained by integrating the first weight, the second weight, the first key degree, the second key degree and the word similarity according to the following formula:

[0081]

[0082] in, represents the first similarity between the i-th first subtext and the j-th second subtext, represents the number of first words corresponding to the i-th first subtext; represents the number of second words corresponding to the j-th second subtext; represents the word similarity between the tth first word of the i-th first subtext and the hth second word of the j-th second subtext, represents the first key degree corresponding to the tth first word of the i-th first subtext, represents the first weight corresponding to the tth first word of the i-th first subtext, represents the second key degree corresponding to the hth second word of the jth second subtext, represents the second weight corresponding to the hth second word of the jth second subtext, represents the similarity between the mth first word and the nth first word of the i-th first subtext, represents the first key degree corresponding to the mth first word of the ith first subtext, represents the first weight corresponding to the mth first word of the ith first subtext, Indicates the first key degree corresponding to the nth first word of the ith first subtext, represents the first weight corresponding to the nth first word of the ith first subtext, represents the similarity between the kth second word and the qth second word of the jth second subtext, represents the second key degree corresponding to the kth second word of the jth second subtext, represents the second weight corresponding to the kth second word of the jth second subtext, The second key degree corresponding to the qth second word of the jth second subtext, Indicates the second weight corresponding to the qth second word of the jth second subtext.

[0083] Exemplarily, according to the above formula, the first weight, the second weight, the first key degree and the second key degree are integrated with the word similarity to obtain the first similarity between the first sub-text and the second sub-text. This introduces the key degree and weight information of the words, which can more accurately reflect the importance of different words in the text, and thus can more accurately determine the first similarity between the first sub-text and the second sub-text.

[0084] Exemplarily, the first similarity is data adjusted according to the word similarity. The adjustment method can be determined according to specific needs and data characteristics, for example, through linear transformation, nonlinear mapping and other methods, so that the final first text similarity can more accurately reflect the similarity between the two subtexts.

[0085] Specifically, traditional text similarity calculations may only consider the surface matching of words or simple semantic similarity, while this method comprehensively considers multiple factors such as word parts, positions, and keyness. These factors can more comprehensively reflect the importance and role of words in the text, so that the calculated similarity can more accurately reflect the true similarity between the two sub-texts. By considering the word parts and positions, the structural and semantic features of the text can be better captured. Words of different parts of speech have different functions in the text, and the position of the word will also affect the focus of its expression. At the same time, the keyness of a word takes into account its rarity and importance in the entire text collection. Combining these factors can provide a deeper understanding of the connotation of the text and improve the ability to capture text features.

[0086] In some embodiments, the data adjustment of the first similarity according to the word similarity to obtain the first text similarity corresponding to the first subtext and the second subtext includes: determining a first associated word corresponding to the first word in the second word when it is close to the first word according to the word similarity, and obtaining a third associated word corresponding to the first associated word from the first word; determining a second associated word corresponding to the second word in the first word when it is close to the second word according to the word similarity, and obtaining a fourth associated word corresponding to the second associated word from the second word; determining a word overlap degree corresponding to the first subtext and the second subtext according to the word similarity; obtaining a first distribution position corresponding to the first associated word from the second subtext, and obtaining the third associated word from the first subtext. a corresponding second distribution position; determining the first word order information corresponding to the first subtext and the second subtext according to the first distribution position and the second distribution position; obtaining the third distribution position corresponding to the second associated word from the first subtext, and obtaining the fourth distribution position corresponding to the fourth associated word from the second subtext; determining the second word order information between the first subtext and the second subtext according to the third distribution position and the fourth distribution position; determining the similarity adjustment parameter between the first subtext and the second subtext according to the first word order information and the second word order information; adjusting the first similarity by a similarity value according to the similarity adjustment parameter to obtain the first text similarity corresponding to the first subtext and the second subtext; wherein the similarity adjustment parameter is obtained according to the following formula:

[0087]

[0088]

[0089] in, represents the similarity adjustment parameter corresponding to the i-th first subtext and the j-th second subtext, represents the number of similar related words between the i-th first subtext and the j-th second subtext, represents the number of the first words corresponding to the i-th first subtext, represents a constant, represents the number of the second words corresponding to the j-th second subtext, represents the first word order information corresponding to the i-th first subtext and the j-th second subtext, Indicates the second word order information corresponding to the j-th second sub-text and the ith first sub-text.

[0090] For example, using the word similarity calculated previously, find words similar to the first word in the second word, define these similar second words as the first associated words corresponding to the first word, and determine the words corresponding to the first associated words in the first word as the third associated words. For example, the first words include "good service" and "average quality", and the second words include "good service" and "not recommended". Through the word similarity, it is judged that "good service" is similar to "good service", so "good service" is the first associated word, and "good service" is the corresponding third associated word.

[0091] For example, similarly, based on the word similarity, a word similar to the second word is found in the first word as the second associated word; then a word corresponding to the second associated word is found in the second word as the fourth associated word.

[0092] Exemplarily, the number of identical or similar words in the first subtext and the second subtext is counted according to the word similarity, so as to determine the word overlap between the two subtexts.

[0093] For example, the occurrence position of the first associated word in the second subtext is determined as the first distribution position; the occurrence position of the third associated word in the first subtext is determined as the second distribution position. Similarly, the occurrence position of the second associated word in the first subtext is determined as the third distribution position; the occurrence position of the fourth associated word in the second subtext is determined as the fourth distribution position.

[0094] Exemplarily, the order relationship between the first associated word and the third associated word in each subtext is analyzed based on the first distribution position and the second distribution position. In order to quantify this word order difference, it is necessary to determine the number of times the first associated word in the second subtext is converted to the order of the third associated word in the first subtext. This can be achieved by a simulated movement method. Specifically, we take the position of the first associated word in the second subtext as the starting point and try to move it to the position of the third associated word in the first subtext. During the movement process, the number of movements is recorded once for each word passed. Thus, the total number of movements from the first associated word to the third associated word can be determined as the first word order information corresponding to the first subtext and the second subtext. This first word order information is a quantitative indicator, which intuitively reflects the degree of word order difference between the two subtexts in the first associated word and the third associated word related parts. The more the number of movements, the greater the word order difference between the two subtexts in this part; the fewer the number of movements, the more similar the word order. The first word order information determined in this way is of great significance in text similarity analysis. It provides us with a new perspective, allowing us to not only focus on the semantic similarity of words, but also to delve deeper into the word order characteristics of the text. Considering the word order information can more accurately judge the similarity between texts, thereby improving accuracy.

[0095] Exemplarily, the order relationship between the second associated word and the fourth associated word in each subtext is analyzed based on the third distribution position and the fourth distribution position. In order to quantify this word order difference, it is necessary to determine the number of times the second associated word in the first subtext is converted to the order of the fourth associated word in the second subtext, so that the total number of times the second associated word is converted to the fourth associated word can be determined as the second word order information corresponding to the second subtext and the first subtext. This second word order information is a quantitative indicator, which intuitively reflects the degree of word order difference between the two subtexts in the relevant parts of the second associated word and the fourth associated word.

[0096] Exemplarily, the first word order information and the second word order information are integrated, and factors such as the consistency and degree of difference of the word order are considered, and the similarity adjustment parameter between the first subtext and the second subtext is determined by the following formula:

[0097]

[0098]

[0099] in, represents the similarity adjustment parameter corresponding to the i-th first subtext and the j-th second subtext, represents the number of similar related words between the i-th first subtext and the j-th second subtext, represents the number of first words corresponding to the i-th first subtext, represents a constant, represents the number of second words corresponding to the j-th second subtext, represents the first word order information corresponding to the i-th first subtext and the j-th second subtext, Indicates the second word order information corresponding to the j-th second subtext and the i-th first subtext.

[0100] Exemplarily, the first word order information and the second word order information respectively reflect the word order relationship of the associated words between the two sub-texts from different angles. By combining these two kinds of word order information, the characteristics of the text in word order can be captured more comprehensively, and then the subtle differences in their expressions can be found by combining the word order information. The similarity adjustment parameter takes word order information into consideration, making the evaluation of text similarity more in line with the actual situation. Factors such as the number of similar associated words, the number of first words, and the number of second words are introduced into the formula. The number of similar associated words reflects the degree of semantic association between the two sub-texts, and the length of the text plays a regulatory role in this degree of association. For example, in a longer text, even if there are a certain number of similar associated words, it cannot be simply assumed that they are highly similar, because only part of the content may be similar. By combining these factors, the semantic similarity between the two sub-texts can be measured more accurately, avoiding misjudgments caused by the influence of text length or the number of associated words.

[0101] For example, the calculation of the similarity adjustment parameter takes into account the consistency and difference of the word order. When the word order of the two sub-texts is relatively consistent, the similarity adjustment parameter will increase accordingly, indicating that their similarity in word order has a positive contribution to the overall similarity; conversely, when the word order difference is large, the similarity adjustment parameter will decrease, thereby reducing the impact of this word order difference on the overall similarity.

[0102] Exemplarily, the calculated similarity adjustment parameter is multiplied by the previously calculated first similarity to obtain the first text similarity corresponding to the first subtext and the second subtext.

[0103] Specifically, the method takes into account the correspondence between word order information and associated words, can adapt to such diverse text expressions, and improve compatibility with different text styles, thereby being able to more accurately obtain the corresponding first text similarity between the first sub-text and the second sub-text.

[0104] Step S103: Obtain relevant information corresponding to the surrounding objects of the target object, and determine a second recommendation parameter according to the relevant information and the target information.

[0105] Exemplarily, the surrounding objects are gas stations around the target object, and then keywords related to the surrounding gas stations are used to search from public social media such as Tik Tok, Weibo, or local life service platforms or community forums, such as the name, address, brand, etc. of the gas station, to obtain relevant information corresponding to the evaluation of the surrounding objects by users on social media.

[0106] Exemplarily, a clustering algorithm such as k-means clustering is used to perform text clustering on relevant information and target information to obtain a target clustering result, and then sentiment classification is performed on each sub-cluster in the target clustering result to obtain the sentiment type corresponding to each sub-cluster, and then each sub-text in each sub-cluster is distinguished as being derived from relevant information or target information, thereby obtaining the number of relevant information and the number of target information in each sub-cluster, and then obtaining the proportion of target information in the sub-cluster, thereby defining the calculation rule of the target score according to the proportion of target information in each sub-cluster and the sentiment type corresponding to the sub-cluster. For example, if the sentiment type is positive and the proportion of target information is high, a higher score is given; if the sentiment type is negative and the proportion of target information is low, a higher score is given, thereby summing the above scores to obtain the target score corresponding to the target object. A mapping relationship between the target score and the second recommendation parameter is established, thereby finding the corresponding second recommendation parameter value according to the target score of the target object, thereby determining the second recommendation parameter corresponding to the target object when competing with the surrounding objects. The second recommendation parameter is used to characterize the user's tendency parameter when selecting services from the target object and the surrounding objects.

[0107] Step S104: determining an associated road section according to the target object and the surrounding objects, and obtaining a historical vehicle flow of the associated road section according to a sensor.

[0108] Exemplarily, accurate geographic location information of the target object and surrounding objects is obtained through a geographic information system (GIS), a global positioning system (GPS) or related map data, and then related road network data is obtained based on the geographic location information, and then associated road sections are determined based on the road network data. The associated road sections are used to represent the corresponding roads that need to pass through the target object or surrounding objects.

[0109] For example, the number of vehicles is counted using image recognition technology using cameras installed on the road, or the historical vehicle flow of the associated road section is obtained by using microwave sensors to detect vehicles by emitting microwave signals and receiving reflected signals.

[0110] Step S105: Obtain corresponding association relationships between the associated road sections according to the historical vehicle flow.

[0111] Exemplarily, the first road where the target object is located is obtained from the associated road section, and then the first data corresponding to the first road is obtained from the historical vehicle flow, and the other directly or indirectly connected second roads of the first road are obtained from the associated road section, and the second data corresponding to the second road is obtained from the historical vehicle flow.

[0112] Exemplarily, a first change trend corresponding to the first data and a second change trend corresponding to the second data are calculated, and data alignment processing is performed according to the first change trend and the second change trend, so as to determine the correlation value between the first road and the second road according to the data alignment result, and then when the correlation value is greater than a preset value, it is determined that the correlation relationship between the first road and the second road is strongly correlated, and when the correlation value is less than or equal to the preset value, it is determined that the correlation relationship between the first road and the second road is weakly correlated.

[0113] In some embodiments, obtaining the corresponding association relationship between the associated road sections based on the historical vehicle flow includes: obtaining the direct road section corresponding to the target object from the associated road sections, and obtaining other road sections except the direct road section from the associated road sections; obtaining first flow data corresponding to the direct road section and second flow data corresponding to the other road sections from the historical vehicle flow; determining first time information corresponding to the first flow data and determining second time information corresponding to the second flow data; determining a data alignment result between the first flow data and the second flow data based on the first flow data, the second flow data combined with the first time information and the second time information; and determining the corresponding association relationship between the direct road section and the other road sections based on the data alignment result.

[0114] For example, the location of the target gas station is determined by finding the road section directly connected to the target gas station from the associated road sections based on the actual connection of the road, that is, the direct-connected road section. The direct-connected road section is used to characterize the road where the target object is located. Among the associated road sections, the directly-connected road sections that have been determined are removed, and the remaining road sections are other road sections.

[0115] For example, the flow data corresponding to the directly connected road section is screened out from the collected historical vehicle flow data, and is defined as the first flow data. Similarly, the flow data corresponding to other road sections is found out from the historical vehicle flow data, namely the second flow data.

[0116] For example, the first traffic data is matched with corresponding time information, which can reflect the vehicle traffic conditions of the directly connected road section at different times. The second traffic data is matched with corresponding time information to understand the changes in vehicle traffic of other road sections at different times.

[0117] Exemplarily, a unified time scale conversion is performed on the first time information and the second time information to ensure that the two are comparable in the time dimension. According to the unified time scale, the first flow data and the second flow data are matched according to the dynamic time warping algorithm, and the flow data at the same time point are correspondingly associated to obtain the data alignment result between the first flow data and the second flow data.

[0118] For example, a statistical method such as correlation analysis, regression analysis, etc. is used to calculate the correlation coefficient between the first flow data and the second flow data through the data alignment result, so as to determine the correlation relationship between the direct-connected road section and the other road section according to the correlation coefficient. If the correlation coefficient is high, it means that there is a strong correlation between the two, which may mean that the flow of vehicles between the direct-connected road section and the other road section is more frequent; otherwise, the correlation is weak.

[0119] For example, if the calculated correlation coefficient is high, close to 1 or -1, it means that there is a strong correlation between the first flow data and the second flow data. When the correlation coefficient is close to 1, it means that the direct-connected section and the other sections are positively correlated, that is, when the vehicle flow of other sections increases, the vehicle flow of the direct-connected section is also likely to increase, which may mean that the flow of vehicles between the direct-connected section and the other sections is more frequent. On the contrary, if the correlation coefficient is close to 0, it means that the correlation between the two is weak. This may mean that the vehicle driving patterns of the direct-connected section and the other sections are relatively independent, and there is less vehicle flow between them.

[0120] In some embodiments, determining the corresponding association relationship between the directly connected road segment and the other road segments based on the data alignment result includes: determining the data alignment quantity between the first flow data and the second flow data based on the data alignment result; when the data alignment quantity is greater than a preset value, determining the corresponding flow difference between the directly connected road segment and the other road segments based on the data alignment quantity in combination with the first flow data and the second flow data; determining a first correlation coefficient corresponding to the first flow data under the association relationship based on the data alignment quantity; determining a second correlation coefficient corresponding to the second flow data under the association relationship based on the data alignment quantity; determining a corresponding association characterization value between the directly connected road segment and the other road segment based on the flow difference in combination with the first correlation coefficient and the second correlation coefficient; and determining the corresponding association relationship between the directly connected road segment and the other road segment based on the correlation characterization value.

[0121] For example, after completing the data alignment operation of the first flow data and the second flow data, the number of data points that can be successfully aligned in the time dimension is counted, and this number is the data alignment number. For example, after time scale unification and data matching, it is found that there are 50 time points at which the first flow data and the second flow data have valid records, then the data alignment number is 50.

[0122] Exemplarily, a numerical value is pre-set as a judgment criterion, namely, a preset value. The calculated data alignment number is compared with the preset value. The setting of the preset value needs to be determined according to the actual situation and data characteristics, for example, it can be comprehensively considered based on factors such as the overall scale of the data and the reliability of the data. If the number of data alignments is greater than the preset value, it means that there are enough data points for further analysis and proceed to the next step; if it is less than or equal to the preset value, it may be impossible to perform accurate correlation analysis due to insufficient data volume, or the correlation may be directly determined to be a weak correlation.

[0123] Exemplarily, when the number of data alignments is greater than a preset value, the first flow data and the second flow data are combined to calculate the flow difference between the directly connected section and other sections. Specifically, for each aligned data point, the difference between the first flow data and the second flow data at that time point is calculated, and then these differences can be processed using statistical methods such as averaging and median to obtain a value that can represent the overall flow difference. For example, for 50 aligned data points, the flow difference at each time point is calculated separately, and then the average of these differences is calculated, and the average value obtained is the flow difference.

[0124] Exemplarily, a first quantity corresponding to the first flow data is obtained, so that a first correlation coefficient corresponding to the first flow data under the correlation relationship is determined according to the ratio between the data alignment quantity and the first quantity. Similarly, a second quantity corresponding to the second flow data is obtained, so that a second correlation coefficient corresponding to the second flow data under the correlation relationship is determined according to the ratio between the data alignment quantity and the second quantity.

[0125] Exemplarily, the first number and the second number are summed to obtain a third number, and then the second number and the third number are ratio-calculated to obtain a first ratio, and the first number and the third number are ratio-calculated to obtain a second ratio.

[0126] Exemplarily, the first ratio and the first correlation coefficient are multiplied to obtain the first result, and the second ratio and the second correlation coefficient are multiplied to obtain the second result, so that the first result and the second result are summed to obtain the target correlation coefficient, and the flow difference and the target correlation coefficient are multiplied to obtain the corresponding correlation characterization value between the direct section and the other section. The correlation characterization value comprehensively considers the flow difference and the influence of the first flow data and the second flow data in the correlation relationship.

[0127] Exemplarily, the association relationship between the direct-connected road section and the other road section is determined based on the calculated association characterization value. For example, different association characterization value ranges can be pre-set, corresponding to different association relationship types. For example, when the association characterization value is greater than a certain higher threshold, it means that there is a strong association between the direct-connected road section and the other road section, and the flow of vehicles between the two road sections is relatively frequent; when the association characterization value is in the middle range, it is a medium association; when the association characterization value is less than a certain lower threshold, it is a weak association.

[0128] Step S106: determine a traffic prediction model according to the historical vehicle traffic, the historical user traffic and the corresponding association relationship between the associated road sections in combination with the first recommended parameter and the second recommended parameter.

[0129] Exemplarily, historical vehicle flow, the corresponding correlation between associated road sections, the first recommended parameter and the second recommended parameter are determined as the input data of the flow prediction model, and historical user flow is determined as the output data of the flow prediction model, and then the model type of the flow prediction model is determined, such as machine learning models (such as decision trees, random forests) and deep learning models (such as convolutional neural networks), etc., and model evaluation indicators such as mean square error, mean absolute error, etc. are determined.

[0130] Exemplarily, the entire data set of historical vehicle traffic, the corresponding association relationship between associated road sections, the first recommended parameter and the second recommended parameter is divided into a training set, a validation set and a test set. Generally speaking, the training set accounts for the majority of the total data (e.g., 70% - 80%) and is used for model parameter learning; the validation set is used to adjust the model's hyperparameters and select the optimal model structure; the test set is used to evaluate the final performance of the model. Thus, the traffic prediction model is trained using the training set, and the error between the model's prediction results and the historical user traffic is minimized by continuously adjusting the model's parameters. During the training process, a cross-validation method can be used to improve the generalization ability of the model.

[0131] For example, based on the evaluation results of the validation set, grid search, random search and other methods are used to find the optimal hyperparameter combination to optimize the model, and then the optimized model is evaluated using the test set to calculate the previously determined evaluation indicators to determine whether the performance of the model meets the requirements. Thus, based on the results of model evaluation and verification, the model with the best performance is selected as the final traffic prediction model.

[0132] Exemplarily, the final traffic prediction model is deployed in actual applications to predict future user traffic of the target object.

[0133] Step S107: Perform user traffic prediction on the target object according to the traffic prediction model to obtain a target prediction result.

[0134] Exemplarily, the current vehicle flow on each of the associated roads is collected, and then the current vehicle flow, the corresponding association relationship between the associated road sections, the first recommended parameter and the second recommended parameter are input into the traffic prediction model to achieve the purpose of user flow prediction for the target object, thereby obtaining the target prediction result.

[0135] Step S108: determining a target state of the target object according to the target prediction result, and performing real-time display according to the target state to obtain a target display result.

[0136] Exemplarily, the state type corresponding to the target object is determined, such as idle, busy, busy, etc., so as to set the flow interval corresponding to each state type, and then compare the obtained target prediction result with the flow interval corresponding to each state type determined previously one by one. According to the comparison result, the current target state of the target object is determined.

[0137] Exemplarily, a display method of the target state on the target screen is designed, and the display method may include text description, icon, color identification, etc., so that the target state is displayed on the target screen corresponding to the target object according to the display method to obtain a target display result.

[0138] See also Figure 2 , Figure 2A real-time display device 200 for gas station information provided in an embodiment of the present application includes a data acquisition module 201, a data clustering module 202, a parameter determination module 203, a flow determination module 204, a relationship determination module 205, a model determination module 206, a data prediction module 207, and a data display module 208, wherein the data acquisition module 201 is used to obtain historical user flow and historical evaluation information of a target object, and obtain target information corresponding to the target object under social media; the data clustering module 202 is used to cluster the historical evaluation information and the target information to obtain a clustering result, and determine a first recommended parameter according to the clustering result; the parameter determination module 203 is used to obtain relevant information corresponding to surrounding objects of the target object, and determine a first recommended parameter according to the relevant information and the target information Determine the second recommended parameter; a traffic determination module 204, used to determine the associated road section according to the target object and the surrounding objects, and obtain the historical vehicle traffic of the associated road section according to the sensor; a relationship determination module 205, used to obtain the corresponding association relationship between the associated road sections according to the historical vehicle traffic; a model determination module 206, used to determine the traffic prediction model according to the historical vehicle traffic, the historical user traffic and the corresponding association relationship between the associated road sections in combination with the first recommended parameter and the second recommended parameter; a data prediction module 207, used to perform user traffic prediction for the target object according to the traffic prediction model to obtain a target prediction result; a data display module 208, used to determine the target state of the target object according to the target prediction result, and perform real-time display according to the target state to obtain a target display result.

[0139] In some implementations, a gas station information real-time display device 200 can be applied to a terminal device.

[0140] It should be noted that, those skilled in the art can clearly understand that, for the sake of convenience and brevity of description, the specific working process of the above-described gas station information real-time display device 200 can refer to the corresponding process in the aforementioned gas station information real-time display method embodiment, and will not be repeated here.

[0141] See also Figure 3 , Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention.

[0142] like Figure 3 As shown, the terminal device 300 includes a processor 301 and a memory 302 , and the processor 301 and the memory 302 are connected via a bus 303 , such as an I2C (Inter-integrated Circuit) bus.

[0143] Specifically, the processor 301 is used to provide computing and control capabilities to support the operation of the entire terminal device. The processor 301 can be a central processing unit (CPU), and the processor 301 can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0144] Specifically, the memory 302 may be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a mobile hard disk.

[0145] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the embodiment of the present invention, and does not constitute a limitation on the terminal device to which the embodiment of the present invention is applied. The specific server may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0146] The processor is used to run a computer program stored in the memory, and implement any one of the gas station information real-time display methods provided by the embodiments of the present invention when executing the computer program.

[0147] In one embodiment, the processor is used to run a computer program stored in the memory, and implements the following steps when executing the computer program:

[0148] Obtain the historical user traffic and historical evaluation information of the target object, and obtain the target information corresponding to the target object under social media;

[0149] Clustering the historical evaluation information and the target information to obtain a clustering result, and determining a first recommendation parameter according to the clustering result;

[0150] Obtaining relevant information corresponding to surrounding objects of the target object, and determining a second recommendation parameter according to the relevant information and the target information;

[0151] Determine an associated road section according to the target object and the surrounding objects, and obtain a historical vehicle flow of the associated road section according to a sensor;

[0152] Obtaining corresponding association relationships between the associated road sections according to the historical vehicle flow;

[0153] Determine a traffic prediction model according to the historical vehicle traffic, the historical user traffic, and the corresponding association relationship between the associated road sections in combination with the first recommended parameter and the second recommended parameter;

[0154] Performing user traffic prediction on the target object according to the traffic prediction model to obtain a target prediction result;

[0155] The target state of the target object is determined according to the target prediction result, and a target display result is obtained by performing real-time display according to the target state.

[0156] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the terminal device described above can refer to the corresponding process in the aforementioned embodiment of the real-time display method of gas station information, and will not be repeated here.

[0157] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any method for real-time display of gas station information provided in the description of the embodiment of the present invention.

[0158] The storage medium may be an internal storage unit of the terminal device described in the foregoing embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc., equipped on the terminal device.

[0159] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0160] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0161] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above description is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for real-time display of gas station information, characterized in that: The method comprises: Obtain the historical user traffic and historical evaluation information of the target object, and obtain the target information corresponding to the target object under social media; Clustering the historical evaluation information and the target information to obtain a clustering result, and determining a first recommendation parameter according to the clustering result; Obtaining relevant information corresponding to surrounding objects of the target object, and determining a second recommendation parameter according to the relevant information and the target information; Determine an associated road section according to the target object and the surrounding objects, and obtain a historical vehicle flow of the associated road section according to a sensor; Obtaining corresponding association relationships between the associated road sections according to the historical vehicle flow; Determine a traffic prediction model according to the historical vehicle traffic, the historical user traffic, and the corresponding association relationship between the associated road sections in combination with the first recommended parameter and the second recommended parameter; Performing user traffic prediction on the target object according to the traffic prediction model to obtain a target prediction result; The target state of the target object is determined according to the target prediction result, and a target display result is obtained by performing real-time display according to the target state.

2. The method according to claim 1, characterized in that The obtaining of target information corresponding to the target object in social media includes: Obtaining a first position corresponding to the target object, and collecting first information corresponding to the target object in social media according to the first position, wherein the first information is used to represent user-published information corresponding to the first position in social media, and the first information includes first text information and first image information; Performing target association on the first text information to obtain relevant text information associated with the target object, and performing target classification on the first image information to obtain an image description object corresponding to the first image information; Obtaining relevant image information associated with the target object from the first image information according to the image description object; Predicting the position of the first information according to the relevant text information and the relevant image information to obtain relevant position information corresponding to the first information; The first information is screened according to the relevant location information and the first location to obtain the target information corresponding to the target object.

3. The method according to claim 2, characterized in that The performing position prediction on the first information according to the relevant text information and the relevant image information to obtain relevant position information corresponding to the first information includes: Using an image coding layer of a position prediction model to extract image features from the relevant image information to obtain a first feature vector; Using the text encoding layer of the position prediction model to extract text features from the relevant text information to obtain a second feature vector; Determine an information matching value between the relevant image information and the relevant text information according to the first feature vector and the second feature vector using the image-text matching layer of the position prediction model; Using the image position determination layer of the position prediction model to extract position signs of the relevant image information, obtaining image position information corresponding to the relevant image information; Using the text position determination layer of the position prediction model to extract positional signs of the relevant text information, obtaining text position information corresponding to the relevant text information; Using the feature fusion layer of the position prediction model, the image position information and the text position information are fused according to the information matching value to obtain a target fusion feature; The location classification layer of the location prediction model uses the target fusion feature to perform location prediction on the first information to obtain the relevant location information corresponding to the first information.

4. The method according to claim 1, characterized in that: The clustering of the historical evaluation information and the target information to obtain a clustering result includes: Merging the historical evaluation information and the target information to obtain a text set, and performing word segmentation processing on each sub-text information in the text set to obtain associated words corresponding to the sub-text information; Determining the first text similarity corresponding to any two of the sub-text information according to the associated words; Determine the target dimension, and obtain the word representation vector corresponding to the associated word, and determine the word dimension association value corresponding to the associated word under the target dimension according to the word representation vector; Determine the weight information corresponding to the sub-text information under the target dimension according to the word dimension association value; Determine, according to the word dimension association value and the weight information, the text dimension association value corresponding to when the sub-text information belongs to the target dimension; Determine the second text similarity corresponding to any two of the sub-text information according to the text dimension association value; The text set is clustered according to the first text similarity and the second text similarity to obtain the clustering result.

5. The method according to claim 4, characterized in that The determining, according to the associated words, the first text similarity corresponding to any two of the sub-text information includes: Obtaining any two subtext information as a first subtext and a second subtext, and obtaining a first word corresponding to the first subtext and a second word corresponding to the second subtext; Determine a first part of speech and a first word position corresponding to the first word in the first subtext, and determine a first weight corresponding to the first word according to the first part of speech and the first word position; Determine a second part of speech and a second word position corresponding to the second word in the second subtext, and determine a second weight corresponding to the second word according to the second part of speech and the second word position; Determining a first criticality corresponding to the first word according to a first frequency corresponding to the first word in the text collection; Determining a second criticality corresponding to the second word according to a second frequency corresponding to the second word in the text collection; Calculating text similarity based on the first word and the second word to obtain word similarity, and fusing the word similarity based on the first weight, the second weight, the first key degree, and the second key degree to obtain a first similarity between the first subtext and the second subtext; The first similarity is data-adjusted according to the word similarity to obtain the first text similarity corresponding to the first subtext and the second subtext.

6. The method according to claim 5, characterized in that The step of performing data adjustment on the first similarity according to the word similarity to obtain the first text similarity corresponding to the first subtext and the second subtext includes: Determine, according to the word similarity, a first associated word in the second word that is close to the first word and corresponds to the first associated word, and obtain a third associated word corresponding to the first associated word from the first word; Determine, according to the word similarity, a second associated word in the first word that corresponds to the second associated word when the second associated word is close to the first word, and obtain a fourth associated word corresponding to the second associated word from the second word; Determining the degree of overlap of corresponding words between the first subtext and the second subtext according to the word similarity; Obtaining a first distribution position corresponding to the first associated word from the second subtext, and obtaining a second distribution position corresponding to the third associated word from the first subtext; Determining first word order information corresponding to the first subtext and the second subtext according to the first distribution position and the second distribution position; Obtaining a third distribution position corresponding to the second associated word from the first subtext, and obtaining a fourth distribution position corresponding to the fourth associated word from the second subtext; determining second word order information between the first subtext and the second subtext according to the third distribution position and the fourth distribution position; determining a similarity adjustment parameter between the first subtext and the second subtext according to the first word order information and the second word order information; Adjusting the first similarity by a similarity value according to the similarity adjustment parameter to obtain the first text similarity corresponding to the first subtext and the second subtext; The similarity adjustment parameter is obtained according to the following formula: in, represents the similarity adjustment parameter corresponding to the i-th first subtext and the j-th second subtext, represents the number of similar related words between the i-th first subtext and the j-th second subtext, represents the number of the first words corresponding to the i-th first subtext, represents a constant, represents the number of the second words corresponding to the j-th second subtext, represents the first word order information corresponding to the i-th first subtext and the j-th second subtext, Indicates the second word order information corresponding to the j-th second sub-text and the ith first sub-text.

7. The method according to claim 1, characterized in that The obtaining the corresponding association relationship between the associated road sections according to the historical vehicle flow includes: Obtaining a directly connected road segment corresponding to the target object from the associated road segments, and obtaining other road segments except the directly connected road segment from the associated road segments; Obtaining first traffic data corresponding to the directly connected road section and second traffic data corresponding to the other road sections from the historical vehicle traffic; Determine first time information corresponding to the first flow data and determine second time information corresponding to the second flow data; Determine a data alignment result between the first flow data and the second flow data according to the first flow data, the second flow data, the first time information, and the second time information; The corresponding association relationship between the directly connected road segment and the other road segments is determined according to the data alignment result.

8. The method according to claim 7, characterized in that The determining the corresponding association relationship between the directly connected road segment and the other road segments according to the data alignment result includes: Determine the data alignment quantity between the first flow data and the second flow data according to the data alignment result; When the data alignment amount is greater than a preset value, the corresponding flow difference between the directly connected section and the other sections is determined according to the data alignment amount combined with the first flow data and the second flow data; Determine a first correlation coefficient corresponding to the first traffic data under the correlation relationship according to the data alignment quantity; Determine a second correlation coefficient corresponding to the second traffic data under the correlation relationship according to the data alignment quantity; Determine the corresponding correlation representation value between the directly connected road segment and the other road segments according to the flow difference combined with the first correlation coefficient and the second correlation coefficient; The corresponding association relationship between the directly connected road segment and the other road segments is determined according to the association representation value.

9. A real-time display device for gas station information, characterized in that: include: The data acquisition module is used to obtain the historical user traffic and historical evaluation information of the target object, and obtain the target information corresponding to the target object under social media; A data clustering module, used for clustering the historical evaluation information and the target information to obtain a clustering result, and determining a first recommendation parameter according to the clustering result; A parameter determination module, used to obtain relevant information corresponding to the surrounding objects of the target object, and determine a second recommended parameter according to the relevant information and the target information; A traffic determination module, used to determine an associated road section according to the target object and the surrounding objects, and obtain a historical vehicle traffic of the associated road section according to a sensor; A relationship determination module, used for obtaining the corresponding association relationship between the associated road sections according to the historical vehicle flow; A model determination module, configured to determine a traffic prediction model according to the historical vehicle traffic, the historical user traffic and the corresponding association relationship between the associated road sections in combination with the first recommended parameter and the second recommended parameter; A data prediction module, used to perform user traffic prediction on the target object according to the traffic prediction model to obtain a target prediction result; The data display module is used to determine the target state of the target object according to the target prediction result, and to display the target state in real time to obtain a target display result.

10. A terminal device, characterized in that: The terminal device includes a processor and a memory; The memory is used to store computer programs; The processor is used to execute the computer program and implement the real-time display method of gas station information according to any one of claims 1 to 8 when executing the computer program.

Images