Gas station data integration method and device based on big data and terminal equipment

Abstract

The embodiment of the invention provides a gas station data integration method and device based on big data and terminal equipment, and belongs to the technical field of data analysis. The method comprises the steps of determining second attribute data related to configuration according to first attribute data and first feature data; obtaining a similar date of the historical date according to the second attribute data; third attribute data and second feature data on similar dates are obtained; combining the first feature data and the second feature data to obtain third feature data, and performing difference calculation on the third feature data to obtain a target difference; performing data adjustment on the third feature data according to the target difference value to obtain fourth feature data; obtaining associated sub-data of the fourth feature data, and performing exception screening on the third feature data according to the associated sub-data to obtain target feature data; performing anomaly identification on the second attribute data and the third attribute data to obtain target attribute data; and performing data integration according to the target feature data and the target attribute data to obtain an integration result.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a gas station data integration method, device and terminal equipment based on big data. Background Art

[0002] In the operation of traditional gas stations, data processing usually involves multiple dimensions, including equilibrium index data, attribute data, etc. These data are usually scattered in different collection systems, lacking unified standards and integration methods, resulting in serious data island phenomenon. Although existing technologies can integrate data from different collection systems into a unified platform (such as a digital large-screen platform) for display, these technologies often simply bring together data from multiple dimensions and fail to provide in-depth data correlation analysis, making it impossible to quickly and accurately extract effective information from the data integration results to prompt users, thereby reducing the operating efficiency of gas stations. Summary of the invention

[0003] The main purpose of the embodiments of the present invention is to provide a gas station data integration method, device and terminal equipment based on big data, aiming to solve the problem in the related technology that it is impossible to quickly and accurately extract effective information from the data integration results to prompt users, thereby reducing the operating efficiency of the gas station.

[0004] In a first aspect, an embodiment of the present invention provides a gas station data integration method based on big data, comprising:

[0005] Obtaining first attribute data and first feature data corresponding to a target gas station on a historical date, wherein the first feature data is used to characterize a balance index corresponding to the target gas station, and the balance index is used to characterize the number of items sold at the target gas station;

[0006] Determine second attribute data related to item configuration of the target gas station according to the first attribute data and the first characteristic data;

[0007] Performing information alignment on the historical date according to the second attribute data to obtain a close date corresponding to the historical date;

[0008] Obtain the third attribute data and the second characteristic data corresponding to the target gas station on the similar date;

[0009] The first feature data and the second feature data are combined to obtain third feature data, and a difference calculation is performed on any two first sub-data in the third feature data to obtain a target difference;

[0010] Performing data adjustment on the third characteristic data according to the target difference to obtain fourth characteristic data;

[0011] Obtaining associated sub-data corresponding to any second sub-data in the fourth characteristic data within a target range, and performing abnormal screening on the third characteristic data according to the associated sub-data to obtain target characteristic data corresponding to the target gas station under the historical date and the similar date;

[0012] Performing abnormality identification on the second attribute data and the third attribute data to obtain target attribute data corresponding to the historical date and the similar date;

[0013] Data integration is performed according to the target feature data and the target attribute data to determine a data integration result corresponding to the target gas station.

[0014] In a second aspect, an embodiment of the present invention provides a gas station data integration device based on big data, comprising:

[0015] A data acquisition module, used to obtain first attribute data and first feature data corresponding to a target gas station on a historical date, wherein the first feature data is used to characterize a balance index corresponding to the target gas station, and the balance index is used to characterize the number of items sold at the target gas station;

[0016] A data screening module, used for determining second attribute data related to the configuration of items at the target gas station according to the first attribute data and the first characteristic data;

[0017] A data alignment module, used for performing information alignment on the historical date according to the second attribute data to obtain a close date corresponding to the historical date;

[0018] An information acquisition module, used to obtain the third attribute data and the second characteristic data corresponding to the target gas station on the similar date;

[0019] A difference calculation module, used for combining the first feature data and the second feature data to obtain third feature data, and calculating the difference between any two first sub-data in the third feature data to obtain a target difference;

[0020] A data adjustment module, used for adjusting the third characteristic data according to the target difference to obtain fourth characteristic data;

[0021] an abnormality screening module, used to obtain the associated sub-data corresponding to any second sub-data in the fourth characteristic data within the target range, and perform abnormality screening on the third characteristic data according to the associated sub-data, to obtain the target characteristic data corresponding to the target gas station under the historical date and the similar date;

[0022] An abnormality identification module, used for performing abnormality identification on the second attribute data and the third attribute data, and obtaining target attribute data corresponding to the historical date and the similar date;

[0023] The data integration module is used to perform data integration according to the target feature data and the target attribute data, and determine the data integration result corresponding to the target gas station.

[0024] 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 gas station data integration methods based on big data provided in the specification of the present invention are implemented.

[0025] The embodiment of the present invention provides a gas station data integration method, device and terminal device based on big data, the method comprising: obtaining first attribute data and first feature data corresponding to a target gas station under a historical date; determining second attribute data related to the configuration of an item at the target gas station according to the first attribute data and the first feature data; aligning information on the historical date according to the second attribute data to obtain a similar date corresponding to the historical date; obtaining third attribute data and second feature data corresponding to the target gas station under the similar date; obtaining the third attribute data and second feature data corresponding to the target gas station under the similar date; merging the first feature data and the second feature data to obtain the third feature data Data, and perform difference calculation on any two first sub-data in the third feature data to obtain the target difference; adjust the third feature data according to the target difference to obtain the fourth feature data; obtain the associated sub-data corresponding to any second sub-data in the fourth feature data within the target range, and perform abnormal screening on the third feature data according to the associated sub-data to obtain the target feature data corresponding to the target gas station under the historical date and the similar date, and then perform abnormal identification on the second attribute data and the third attribute data to obtain the target attribute data corresponding to the historical date and the similar date; perform data integration according to the target feature data and the target attribute data to determine the data integration result corresponding to the target gas station. The method can effectively identify abnormal data by obtaining the first feature data and the first attribute data of the target gas station under the historical date, and comparing them with the third attribute data and the second feature data of the similar date. This process helps to screen out more accurate feature data and attribute data, thereby improving the accuracy of data analysis. By aligning information between historical dates and similar dates and identifying abnormal feature data and attribute data, noise and errors in the data can be eliminated to ensure the reliability of the analysis results. The data integration results obtained through data integration can be used to provide a basis for resource allocation and business adjustments of gas stations. As a result, the target gas station can obtain a more accurate balance index, further optimize the operation strategy, and improve the scientific nature of decision support. This method also solves the problem in related technologies that it is impossible to quickly and accurately extract effective information from the data integration results to prompt users, thereby reducing the operating efficiency of gas stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] 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.

[0027] Figure 1 A flow chart of a gas station data integration method based on big data provided by an embodiment of the present invention;

[0028] Figure 2 A schematic diagram of the module structure of a gas station data integration device based on big data provided by an embodiment of the present invention;

[0029] Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] 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.

[0031] 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.

[0032] 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.

[0033] The embodiment of the present invention provides a gas station data integration method, device and terminal device based on big data. The gas station data integration method based on big data 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.

[0034] 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.

[0035] Please refer to Figure 1 , Figure 1 A flow chart of a gas station data integration method based on big data provided in an embodiment of the present invention.

[0036] like Figure 1 As shown, the gas station data integration method based on big data includes steps S101 to S109.

[0037] Step S101, obtaining first attribute data and first characteristic data corresponding to a target gas station on a historical date, wherein the first characteristic data is used to characterize a balance index corresponding to the target gas station, and the balance index is used to characterize the number of items sold at the target gas station.

[0038] Exemplarily, the database or data management system of the target gas station is accessed. In the data management system, the query function is used to obtain the first attribute data corresponding to the target gas station under the historical date according to the name corresponding to the target gas station as the screening condition. The first attribute data includes but is not limited to the geographical location, facilities, service type, holiday information corresponding to the historical date, passenger flow or vehicle flow information, etc. of the gas station. Similarly, in the data management system, the name and historical date of the target gas station are used as the screening conditions to find the first characteristic data of the corresponding time period. The first characteristic data is used to characterize the balance index corresponding to the target gas station. The balance index is used to characterize the number of items sold at the target gas station, wherein the first characteristic data includes but is not limited to the number of items corresponding to the oil type or non-oil type that have been sold.

[0039] Step S102: Determine second attribute data related to item configuration of the target gas station according to the first attribute data and the first characteristic data.

[0040] Exemplarily, a preliminary check is performed on the first attribute data and the first characteristic data obtained to remove outliers, missing values ​​and duplicate data to ensure the accuracy and consistency of the data. A statistical analysis is performed on the first characteristic data to check the changing trend of the equilibrium index, and then the distribution of the equilibrium index data is analyzed to understand the peaks and troughs of the equilibrium index. Based on expert experience, attributes that may be related to the equilibrium index are preliminarily screened out from the first attribute data, such as the geographical location of the gas station, weather conditions, holidays, promotional activities, competitor conditions, road traffic conditions, etc., so as to use statistical methods (such as correlation coefficients, regression analysis, etc.) to analyze the correlation between each attribute in the first attribute data and the first characteristic data, and screen out attribute data that is highly correlated with the equilibrium index to form the second attribute data.

[0041] In some embodiments, determining the second attribute data related to the target gas station and the item configuration based on the first attribute data and the first characteristic data includes: obtaining first related data corresponding to the first attribute data under a first data type and second related data corresponding to the first attribute data under a second data type, the first data type being discrete data and the second data type being categorical data; calculating a first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data; calculating a second degree of association between the third sub-data and the first characteristic data, and determining a third degree of association between the fourth sub-data and the first characteristic data based on the first degree of association and the second degree of association; and screening the first attribute data based on the second degree of association and the third degree of association to obtain the second attribute data related to the target gas station and the item configuration.

[0042] Exemplarily, discrete data such as vehicle flow data and passenger flow data are extracted from the first attribute data, and these data are usually expressed as a certain value. Vehicle flow data refers to the number of vehicles passing near a gas station per unit time, and is usually expressed in the form of discrete values. Classified data are extracted from the first attribute data, such as the weather type (sunny, rainy, snowy, etc.) or time period (weekdays, weekends, holidays, etc.) in the area where the gas station is located. These attributes have clear classification labels.

[0043] Exemplarily, a suitable statistical method (such as chi-square test, mutual information, correlation coefficient, etc.) is selected to calculate the correlation between the first related data (discrete type) and the second related data (classified type). For example, the chi-square test can be used to analyze the correlation between "whether there is a convenience store" and "traffic flow data".

[0044] For example, suppose you want to analyze the relationship between "whether there is a convenience store" (categorical) and "traffic flow data" (discrete). You can use "whether there is a convenience store" as a categorical variable, group the traffic flow data (such as low, medium, and high flow), and then apply the chi-square test to determine whether there is a significant association between the two.

[0045] Exemplarily, the correlation between the third sub-data and the first feature data is calculated using a statistical algorithm such as a Pearson correlation coefficient, a Spearman correlation coefficient, etc. For example, the linear correlation between the "discrete vehicle flow data" and the "balance index" can be calculated.

[0046] Exemplarily, the third degree of association between the fourth sub-data (classified data) and the first feature data is comprehensively calculated based on the first degree of association and the second degree of association.

[0047] For example, assuming that the correlation between "whether there is a convenience store" and "traffic flow data", as well as the correlation between "traffic flow data" and "balance index", is known, the indirect correlation between "whether there is a convenience store" and "balance index", that is, the third correlation, can be calculated by multiplication or weighted sum.

[0048] For example, a reasonable screening threshold is set according to the magnitude of the second correlation and the third correlation (e.g., the correlation is greater than a certain standard value). All attributes in the first attribute data are screened, and those second attribute data with a high correlation with the balance index are retained. For example, attributes that have a significant impact on the balance index are screened, such as "holidays" and "promotional activities".

[0049] Specifically, by analyzing the correlation between discrete data and categorical data and the balance index, we can more comprehensively understand the impact of different attributes on the balance index. By setting the correlation threshold, we can filter out the second attribute data that really has a significant impact on the balance index, avoid the interference of invalid or secondary attributes, and improve the accuracy of the analysis.

[0050] In some embodiments, the calculation of the first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data includes: calculating the distance information between any two of the third sub-data in the first related data to obtain a target distance, and determining a first matrix based on the target distance; calculating the degree of difference between any two of the fourth sub-data in the second related data to obtain a target difference value, and determining a second matrix based on the first matrix and the target difference value; performing data fusion based on the first matrix and the second matrix to obtain the target distance between the third sub-data and the fourth sub-data; and determining the first degree of association between the second related data and the first related data based on the target distance.

[0051] Exemplarily, a suitable distance measurement method is selected according to the characteristics of the first related data, such as Euclidean distance, Manhattan distance, etc. For any two third sub-data in the first related data, the distance between them is calculated, and these distance values ​​are organized into a matrix, namely, the first matrix.

[0052] Exemplarily, a suitable dissimilarity measurement method is selected according to the characteristics of the second related data, such as Jaccard distance, Hamming distance, etc. For any two fourth sub-data in the second related data, the dissimilarity between them is calculated, and these dissimilarity values ​​are organized into a matrix, and then the first matrix and the dissimilarity matrix are comprehensively considered, and a second matrix that comprehensively reflects the information of the two matrices may be generated through weighted averaging, direct merging or other mathematical operations.

[0053] Exemplarily, an appropriate fusion method, such as multi-view learning, feature-level fusion, etc., is used to fuse the information of the first matrix and the second matrix, and the target distance between the third sub-data and the fourth sub-data is calculated. According to the target distance obtained after fusion, a correlation calculation method, such as an inverse distance weighting method, similarity conversion, etc., is used to quantify the correlation degree between the second related data and the first related data, and then obtain the first correlation degree between the second related data and the first related data.

[0054] Specifically, by integrating the association information of discrete and categorical data, a more comprehensive perspective is provided to evaluate the association between attributes. By combining different distance and dissimilarity metrics, the intrinsic relationship between data can be captured more accurately, reducing the deviation that may be caused by a single indicator. The target distance and association obtained by integrating multi-source information can provide more accurate data support for the operational decisions of gas stations.

[0055] Step S103: align the historical date information according to the second attribute data to obtain a close date corresponding to the historical date.

[0056] For example, according to the second attribute data selected in the previous step, these attributes usually include features that have a significant impact on the equilibrium index, such as "holidays", "weather type", "promotional activities", "traffic flow information", etc. Ensure that each historical date corresponds to a complete record of these second attribute data, such as the weather type of each date, whether it is a holiday, etc.

[0057] Exemplarily, according to the characteristics of the second attribute data, a suitable similarity measurement method is selected. For example, for categorical attributes (such as weather type, whether it is a holiday), Jaccard similarity, Hamming distance, etc. can be used; for numerical attributes (such as the intensity of promotional activities), Euclidean distance, Manhattan distance, etc. can be used. If the second attribute data contains multiple types of attributes, you can consider combining multiple similarity measurement methods, such as by weighted average or product, to generate a comprehensive similarity score. For each historical date, calculate the similarity between it and all other historical dates to generate a similarity matrix. Each element in the matrix represents the similarity score between two dates.

[0058] Assume that a historical date is D1, and its second attribute data is {holiday: yes, weather type: sunny, promotion: moderate}, and the second attribute data of another date D2 is {holiday: no, weather type: sunny, promotion: moderate}. The similarity of D1 and D2 on each attribute can be calculated separately, and then these similarities are combined to get a total score.

[0059] Exemplarily, according to business needs and actual conditions, a similarity threshold is set to determine whether two dates are close enough. For example, dates with a similarity greater than 0.8 can be set as similar dates, and then the above similarity matrix is ​​used to filter out dates corresponding to each historical date whose similarity exceeds the set threshold. These dates are similar dates to the historical dates. Assume that in the similarity matrix of historical date D1, the similarities of D2 and D3 are 0.85 and 0.9 respectively, and the similarity of D4 is 0.7. Then D2 and D3 are selected as similar dates to D1, and D4 is excluded.

[0060] In some embodiments, the information alignment of the historical date according to the second attribute data to obtain a similar date corresponding to the historical date includes: obtaining a first inherent attribute corresponding to the historical date from the second attribute data, and determining a first entity representation corresponding to the historical date according to the first inherent attribute; obtaining a second inherent attribute corresponding to a date to be judged other than the historical date from the second attribute data, and determining a second entity representation corresponding to the date to be judged according to the second inherent attribute; obtaining a first variable attribute corresponding to the historical date from the second attribute data, and performing entity enhancement on the first entity representation according to the first variable attribute to obtain a third entity representation; obtaining a second variable attribute corresponding to the date to be judged from the second attribute data, and performing entity enhancement on the second entity representation according to the second variable attribute to obtain a fourth entity representation; determining a date similarity between the historical date and the date to be judged according to the third entity representation and the fourth entity representation; and filtering out the similar date corresponding to the historical date from the date to be judged according to the date similarity.

[0061] Exemplarily, inherent attributes related to the historical date are extracted from the second attribute data. These attributes are usually factors that do not change over time, such as the geographical location of the gas station, surrounding facilities, etc. The extracted first inherent attributes are combined into a vector or matrix as the first entity representation of the historical date. For example, the attributes such as the geographical location and surrounding facilities can be digitized to generate a first entity representation representing the historical date. The inherent attributes related to the date to be judged are extracted from the second attribute data, which also include the digitization of the attributes such as the geographical location and surrounding facilities. The extracted second inherent attributes are combined into a vector or matrix as the second entity representation of the date to be judged. For example, the attributes such as the geographical location and surrounding facilities of the date to be judged are digitized to generate a second entity representation representing the date.

[0062] Exemplarily, the changing attributes related to the historical date are extracted from the second attribute data, and these attributes are usually dynamic factors that change over time, such as weather type, promotion intensity, traffic flow information, etc. The first entity representation is enhanced according to the extracted first changing attributes. For example, the changing attributes such as weather type and promotion intensity can be combined with the first inherent attribute to generate an enhanced entity representation containing dynamic information to obtain the third entity representation.

[0063] Exemplarily, the variable attributes related to the date to be determined, such as weather type and promotion intensity, are extracted from the second attribute data. The second entity representation is enhanced according to the extracted second variable attributes. For example, the variable attributes such as weather type and promotion intensity are combined with the second inherent attributes to generate an enhanced entity representation containing dynamic information to obtain a fourth entity representation.

[0064] Exemplarily, an appropriate distance or similarity measurement method (such as Euclidean distance, cosine similarity, Jaccard similarity, etc.) is used to calculate the similarity between the third entity representation and the fourth entity representation. For example, the cosine similarity of the two enhanced entity representations can be calculated to obtain a numerical value representing the similarity between the historical date and the date to be judged. Therefore, according to business needs and actual conditions, a similarity threshold is set to determine which dates to be judged are close enough to the historical dates. For example, dates with a similarity greater than 0.8 are set as similar dates. Therefore, based on the calculated date similarity, the dates to be judged whose similarity with the historical dates exceeds the set threshold are screened out as similar dates to the historical dates.

[0065] In some embodiments, the entity enhancement of the first entity representation according to the first change attribute to obtain the third entity representation includes: using the relationship-aware network of the entity enhancement model to obtain first enhancement information obtained by enhancing the first entity representation in the entity relationship; using the attribute-aware network of the entity enhancement model to calculate the second enhancement information between the first change attribute and the first entity representation; using the information merging network of the entity enhancement model to merge the first enhancement information and the second enhancement information to obtain target enhancement information; and using the information enhancement attention network of the entity enhancement model to entity enhance the first entity representation according to the target enhancement information to obtain the third entity representation.

[0066] Exemplarily, the first enhanced information is obtained by using a relationship-aware network, which is used to capture the associations between entities and enhance the entity representation based on these relationships. The first entity representation is then input into the relationship-aware network, and the first enhanced information is generated by analyzing the relationships between entities (such as the relationship between a gas station and surrounding facilities, the relationship between a geographical location and traffic conditions, etc.). Assuming that the first entity representation is the geographical location of a gas station and the surrounding traffic conditions, the relationship-aware network can analyze the relationship between the gas station and nearby convenience stores and gas stations, and generate enhanced information based on these relationships.

[0067] Exemplarily, the second enhanced information is calculated using an attribute-aware network, which is used to capture the attribute characteristics of an entity and calculate the correlation between the attribute and the entity representation. The first variable attribute (such as weather type, intensity of promotional activities, etc.) and the first entity representation are input into the attribute-aware network, and the impact of these variable attributes on the entity representation is calculated to generate the second enhanced information. Assuming that the first variable attribute is the weather type (sunny, rainy, snowy), the attribute-aware network can calculate the impact of different weather types on the balance index of the gas station and generate enhanced information based on the weather attribute.

[0068] Exemplarily, the first enhancement information and the second enhancement information are merged using an information merging network, which is used to fuse the enhancement information from different networks to generate comprehensive target enhancement information, and then the first enhancement information (from the relationship perception network) and the second enhancement information (from the attribute perception network) are input into the information merging network, and the target enhancement information is generated by weighted averaging, splicing or other fusion methods. Assuming that the first enhancement information is based on the relationship between the gas station and the surrounding facilities, and the second enhancement information is based on the influence of weather attributes, the information merging network can fuse these two parts of information to generate target enhancement information.

[0069] Exemplarily, an information-enhanced attention network is used to perform entity enhancement. The information-enhanced attention network is used to dynamically adjust the weight of the entity representation according to the target enhancement information to achieve the enhancement of the entity representation. The target enhancement information and the first entity representation are input into the information-enhanced attention network. The network dynamically weights the first entity representation according to the key features in the target enhancement information to generate a third entity representation. Assuming that the target enhancement information indicates that the weather type has a greater impact on the equilibrium index, the information-enhanced attention network will enhance the weight of the weather attribute in the entity representation and generate an enhanced entity representation that includes the weather impact.

[0070] Specifically, through the collaborative work of the relationship-aware network and the attribute-aware network, the relationship and attribute characteristics between entities can be captured simultaneously, realizing multi-level information fusion. The information merging network fuses the enhanced information from different networks to generate more comprehensive target enhanced information and improve the expressiveness of entity representation. The information-enhanced attention network dynamically adjusts the weight of entity representation according to the key features in the target enhanced information, so that the enhanced entity representation can more accurately reflect important information. Through the dynamic attention mechanism, the model can flexibly adapt to different business scenarios and changing attributes, and enhance the generalization ability of the model. Therefore, through the collaborative work of multiple networks, the static and dynamic attributes of entities can be fully captured, and the expressiveness of entity representation can be improved. The information-enhanced attention network can efficiently utilize the target enhanced information, generate more accurate enhanced entity representation, and improve the prediction and analysis capabilities of the model. This process not only improves the expressiveness of entity representation and the efficiency of information utilization, but also provides scientific data support for the optimization of gas station operation strategies, helping to achieve more efficient management and decision-making.

[0071] Step S104: Obtain the third attribute data and the second characteristic data corresponding to the target gas station on the similar date.

[0072] Exemplarily, the third attribute data and the second feature data corresponding to the target gas station on similar dates are obtained from the database, wherein the third attribute data and the second attribute data have the same dimension.

[0073] Step S105: merge the first feature data and the second feature data to obtain third feature data, and perform difference calculation on any two first sub-data in the third feature data to obtain a target difference.

[0074] Exemplarily, the first feature data and the second feature data are combined to obtain the third feature data. Assuming that the first feature data is the equilibrium index data of a historical date (January 1, 2023), and the second feature data is the equilibrium index data of a similar date (January 8, 2023, January 15), the third feature data is obtained after the combination. Two first sub-data are randomly selected from the third feature data, and their difference is calculated to obtain the target difference.

[0075] Step S106: adjusting the third characteristic data according to the target difference to obtain fourth characteristic data.

[0076] Exemplarily, the third characteristic data is adjusted according to the target difference to obtain the fourth characteristic data. Assume that the balance index data of January 1 and January 8, 2023 are selected, and the balance index difference of these two days is calculated. Thus, the third characteristic data is adjusted according to the target difference to obtain the fourth characteristic data. Assuming that the target difference is less than or equal to the preset value, it means that the balance indexes of the two dates are similar or equal, and then the two first sub-data can be modified to the same data or the average value between the two; when the target difference is greater than the preset value, it means that the balance indexes of the two dates are greatly different, and no adjustment is required, thereby obtaining the adjusted fourth characteristic data.

[0077] Step S107, obtaining the associated sub-data corresponding to any second sub-data in the fourth characteristic data within the target range, and performing abnormal screening on the third characteristic data according to the associated sub-data to obtain the target characteristic data corresponding to the target gas station on the historical date and the similar date.

[0078] Exemplarily, a second sub-data is randomly selected from the fourth characteristic data to determine its associated sub-data within the target range. Assuming the target range is the nearest 7 data, the 7 data closest to the second sub-data are obtained from the fourth characteristic data and determined as the associated sub-data.

[0079] Exemplarily, statistical feature analysis is performed on the associated sub-data to obtain target statistical features corresponding to the associated sub-data, and then the gap between the target statistical features is calculated, and the second sub-data corresponding to the target statistical feature gap is determined as abnormal data, and then the abnormal data is determined as the target feature data corresponding to the target gas station on historical dates and similar dates.

[0080] In some embodiments, the performing abnormal screening on the third characteristic data according to the associated sub-data to obtain the target characteristic data corresponding to the target gas station at the historical date and the similar date includes: performing information entropy calculation on the associated sub-data to obtain entropy information, and calculating variance information corresponding to the associated sub-data; determining a target characterization value corresponding to the second sub-data within the target range according to the entropy information and the variance information; performing abnormal screening on the third characteristic data according to the target characterization value to obtain the target characteristic data corresponding to the target gas station at the historical date and the similar date; wherein the target characterization value is obtained according to the following formula:

[0081]

[0082] in, represents the weight corresponding to the entropy value information, represents the weight corresponding to the variance information, m represents the target range, represents the frequency information corresponding to the j-th second sub-data, represents the data corresponding to the j-th second sub-data, represents the data corresponding to the t-th second sub-data, and e represents a natural constant.

[0083] Exemplarily, information entropy calculation is performed on the associated sub-data to obtain entropy information, and variance calculation is performed on the associated sub-data to obtain variance information, and then the entropy information and variance information are fused according to the following formula to obtain the target representation value corresponding to the second sub-data:

[0084]

[0085] in, Represents the weight corresponding to the entropy information, represents the weight corresponding to the variance information, m represents the target range, represents the frequency information corresponding to the j-th second sub-data, represents the data corresponding to the j-th second sub-data, represents the data corresponding to the t-th second sub-data, and e represents a natural constant. The target characterization value is used to characterize the fluctuation information corresponding to the second sub-data.

[0086] Exemplarily, a target threshold is determined, and then the target threshold is compared with the target characterization value. When the target characterization value is greater than the target threshold, the second sub-data is determined as abnormal data, otherwise the second sub-data is determined as normal data. Thus, the abnormal data is removed from the fourth characteristic data, so as to obtain the target characteristic data of the target gas station under the historical date and the similar date.

[0087] Step S108: perform anomaly identification on the second attribute data and the third attribute data to obtain target attribute data corresponding to the historical date and the similar date.

[0088] Exemplarily, anomaly identification methods such as statistical methods (such as Z-score, IQR), machine learning methods (such as isolation forest, LOF), etc. are used. According to the selected anomaly identification method, a suitable anomaly threshold is set to determine whether the second attribute data and the third attribute data are abnormal, so as to remove the abnormal data from the second attribute data and the third attribute data to obtain the target attribute data corresponding to the target gas station on the historical date and the similar date.

[0089] Step S109: Perform data integration based on the target feature data and the target attribute data to determine the data integration result corresponding to the target gas station.

[0090] Exemplarily, the target feature data and target attribute data are data related to similar dates, and then the target feature data and target attribute data are integrated and stored according to similar dates, so that after the target feature data and target attribute data are stored in the database according to similar dates, the target feature data and target attribute data are displayed on the interface, so as to obtain the data integration result corresponding to the target gas station. That is, the data integration result is used to characterize the interface display result corresponding to the target feature data and target attribute data with similar information.

[0091] For example, choose a suitable display platform, such as a web interface, mobile application, data visualization tool, etc., to display the integrated data, and then display the target feature data and target attribute data in the form of charts, tables, etc., so that users can view and analyze them intuitively. For example, use Tableau or Power BI to create an interactive dashboard to display the balance index trend chart and attribute data chart of the target feature data and target attribute data.

[0092] In some embodiments, after obtaining the data integration result, the method further includes: determining a target prediction date corresponding to the target user, and obtaining a third inherent attribute and a third change attribute corresponding to the target prediction date; obtaining a target associated date corresponding to the target prediction date from the historical date according to the third inherent attribute and the third change attribute; obtaining prediction feature data corresponding to the target prediction date from the data integration result according to the target associated date; and determining a resource scheduling strategy corresponding to the target prediction date according to the prediction feature data.

[0093] For example, the target prediction date corresponding to the target user is determined, and then the third inherent attribute and the third variable attribute corresponding to the target prediction date are determined according to the historical information, the third inherent attribute such as geographical location, gas station scale, oil type, etc. The third variable attribute such as weather forecast, holidays, promotional activities, etc.

[0094] Exemplarily, dates with similar attributes to the target prediction date are screened out from historical dates as target associated dates, and then based on the target associated dates, corresponding equilibrium index data are extracted from the data integration results, and then the prediction model corresponding to the equilibrium index data is determined, such as time series analysis, regression analysis, etc., and prediction is performed based on the equilibrium index data of the target associated date to obtain prediction feature data.

[0095] For example, according to the predicted characteristic data, a resource scheduling strategy is formulated, such as oil inventory adjustment, personnel arrangement, equipment preparation, etc. Assuming that the predicted characteristic data is 5,000 liters, the resource scheduling strategy is formulated to reserve 5,000 liters of oil in advance and arrange sufficient refuelers and equipment.

[0096] Specifically, through the analysis of inherent attributes and changing attributes, the equilibrium index data of the target forecast date can be predicted more accurately, improving the accuracy of the forecast. Based on the historical equilibrium index data of the target associated date, combined with the forecast model, a more scientific and data-driven equilibrium index forecast can be achieved. According to the forecast feature data, the resource scheduling strategy can be dynamically adjusted to ensure the reasonable allocation of oil inventory, personnel and equipment. Therefore, through accurate resource scheduling strategies, unnecessary inventory backlogs and resource waste can be reduced, and operating costs can be reduced.

[0097] In some embodiments, obtaining the target associated date corresponding to the target predicted date from the historical date according to the third inherent attribute and the third changing attribute includes: determining a first attenuation factor and a second attenuation factor, and determining a first similarity between the target predicted date and the historical date using the third inherent attribute according to the first attenuation factor and the second attenuation factor; determining a data association relationship according to the third changing attribute and a fourth changing attribute corresponding to the historical date to obtain a second similarity; and fusing the first similarity and the second similarity to determine the target associated date corresponding to the target predicted date.

[0098] Exemplarily, based on historical data and business needs, a first attenuation factor is determined to measure the degree of attenuation of the third inherent attribute to the similarity of historical dates. Assuming the first attenuation factor is 0.8, it means that the contribution of the third inherent attribute to the similarity of historical dates decays faster over time, that is, the greater the time interval between the historical date and the target prediction date, the lower the similarity. Based on historical data and business needs, a second attenuation factor is determined to measure the degree of attenuation of the third variable attribute to the similarity of historical dates. Assuming the second attenuation factor is 0.6, it means that the contribution of the third variable attribute to the similarity of historical dates decays slowly over time, that is, even if the time interval is large, the third variable attribute still contributes to the similarity.

[0099] Exemplarily, the first similarity is calculated based on the third intrinsic attribute and the intrinsic attribute corresponding to each similar date in the data integration result using weighted Euclidean distance and cosine similarity. The third change attribute corresponding to the target forecast date is extracted, such as weather, holidays, promotional activities, etc. The change attribute corresponding to the historical date is extracted, and then the second similarity is calculated based on the third change attribute and the change attribute corresponding to each similar date in the data integration result using weighted Euclidean distance, cosine similarity and other methods.

[0100] Exemplarily, the first similarity and the second similarity are weightedly fused according to the first attenuation factor and the second attenuation factor to obtain the target similarity, and then the date with the highest target similarity is screened from the data integration results according to the target similarity as the target associated date.

[0101] See also Figure 2 , Figure 2 A gas station data integration device 200 based on big data is provided for an embodiment of the present application. The gas station data integration device 200 based on big data includes a data acquisition module 201, a data screening module 202, a data alignment module 203, an information acquisition module 204, a difference calculation module 205, a data adjustment module 206, an abnormal screening module 207, an abnormal identification module 208, and a data integration module 209, wherein the data acquisition module 201 is used to obtain the first attribute data and the first feature data corresponding to the target gas station under the historical date, the first feature data is used to characterize the balance index corresponding to the target gas station, and the balance index is used to characterize the number of items sold at the target gas station; the data screening module 202 is used to determine the second attribute data related to the item configuration of the target gas station according to the first attribute data and the first feature data; the data alignment module 203 is used to align the historical date with the second attribute data to obtain the similar date corresponding to the historical date; the information acquisition module 204 is used to obtain the target gas station in the corresponding The third attribute data and the second characteristic data corresponding to the recent date; a difference calculation module 205, which is used to merge the first characteristic data and the second characteristic data to obtain the third characteristic data, and calculate the difference between any two first sub-data in the third characteristic data to obtain the target difference; a data adjustment module 206, which is used to adjust the third characteristic data according to the target difference to obtain the fourth characteristic data; an abnormal screening module 207, which is used to obtain the associated sub-data corresponding to any second sub-data in the fourth characteristic data within the target range, and perform abnormal screening on the third characteristic data according to the associated sub-data to obtain the target characteristic data corresponding to the target gas station under the historical date and the recent date; an abnormal identification module 208, which is used to perform abnormal identification on the second attribute data and the third attribute data to obtain the target attribute data corresponding to the historical date and the recent date; a data integration module 209, which is used to perform data integration according to the target characteristic data and the target attribute data to determine the data integration result corresponding to the target gas station.

[0102] In some implementations, during the process of determining the second attribute data related to the object configuration of the target gas station according to the first attribute data and the first feature data, the data screening module 202 performs:

[0103] Obtaining first related data corresponding to the first attribute data under a first data type and second related data corresponding to a second data type, wherein the first data type is discrete data and the second data type is categorical data;

[0104] calculating a first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data;

[0105] calculating a second degree of association between the third sub-data and the first characteristic data, and determining a third degree of association between the fourth sub-data and the first characteristic data according to the first degree of association and the second degree of association;

[0106] The first attribute data is screened according to the second degree of association and the third degree of association to obtain the second attribute data related to the target gas station and the item configuration.

[0107] In some implementations, the data screening module 202 performs, during the process of calculating the first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data:

[0108] Calculate the distance information between any two of the third sub-data in the first related data to obtain a target distance, and determine a first matrix according to the target distance;

[0109] Calculating the degree of difference between any two fourth sub-data in the second related data to obtain a target difference value, and determining a second matrix according to the first matrix and the target difference value;

[0110] Performing data fusion according to the first matrix and the second matrix to obtain a target distance between the third sub-data and the fourth sub-data;

[0111] The first degree of association between the second related data and the first related data is determined according to the target distance.

[0112] In some implementations, the data alignment module 203 performs, during the process of aligning the historical date according to the second attribute data to obtain a similar date corresponding to the historical date:

[0113] Obtaining a first inherent attribute corresponding to the historical date from the second attribute data, and determining a first entity representation corresponding to the historical date according to the first inherent attribute;

[0114] Obtaining a second inherent attribute corresponding to the date to be determined other than the historical date from the second attribute data, and determining a second entity representation corresponding to the date to be determined according to the second inherent attribute;

[0115] Obtaining a first change attribute corresponding to the historical date from the second attribute data, and performing entity enhancement on the first entity representation according to the first change attribute to obtain a third entity representation;

[0116] Obtaining a second change attribute corresponding to the to-be-determined date from the second attribute data, and performing entity enhancement on the second entity representation according to the second change attribute to obtain a fourth entity representation;

[0117] Determine the date similarity between the historical date and the date to be determined according to the third entity representation and the fourth entity representation;

[0118] The close date corresponding to the historical date is selected from the dates to be determined according to the date similarity.

[0119] In some implementations, during the process of performing entity enhancement on the first entity representation according to the first change attribute to obtain the third entity representation, the data alignment module 203 performs:

[0120] Obtaining first enhanced information obtained by enhancing the first entity representation in the entity relationship using a relationship-aware network of an entity enhancement model;

[0121] Calculating second enhancement information between the first changed attribute and the first entity representation using an attribute-aware network of the entity enhancement model;

[0122] Using the information merging network of the entity enhancement model, the first enhancement information and the second enhancement information are merged to obtain target enhancement information;

[0123] The information enhancement attention network of the entity enhancement model is used to perform entity enhancement on the first entity representation according to the target enhancement information to obtain the third entity representation.

[0124] In some implementations, the abnormality screening module 207, in the process of screening the third characteristic data for abnormalities according to the associated sub-data to obtain the target characteristic data corresponding to the target gas station on the historical date and the similar date, performs:

[0125] Performing information entropy calculation on the associated sub-data to obtain entropy value information, and calculating variance information corresponding to the associated sub-data;

[0126] Determine a target representation value corresponding to the second sub-data within the target range according to the entropy value information and the variance information;

[0127] Performing abnormal screening on the third characteristic data according to the target characterization value to obtain the target characteristic data corresponding to the target gas station on the historical date and the similar date;

[0128] The target characterization value is obtained according to the following formula:

[0129]

[0130] in, represents the weight corresponding to the entropy value information, represents the weight corresponding to the variance information, m represents the target range, represents the frequency information corresponding to the j-th second sub-data, represents the data corresponding to the j-th second sub-data, represents the data corresponding to the t-th second sub-data, and e represents a natural constant.

[0131] In some implementations, after obtaining the data integration result, the gas station data integration device 200 based on big data further performs:

[0132] Determine a target prediction date corresponding to the target user, and obtain a third inherent attribute and a third change attribute corresponding to the target prediction date;

[0133] Obtaining a target associated date corresponding to the target forecast date from the historical dates according to the third inherent attribute and the third variable attribute;

[0134] Obtaining prediction feature data corresponding to the target prediction date from the data integration result according to the target association date;

[0135] A resource scheduling strategy corresponding to the target forecast date is determined according to the forecast characteristic data.

[0136] In some implementations, the big data-based gas station data integration device 200, in the process of obtaining the target associated date corresponding to the target predicted date from the historical date according to the third inherent attribute and the third change attribute, performs:

[0137] Determining a first attenuation factor and a second attenuation factor, and determining a first similarity between the target forecast date and the historical date using the third inherent attribute according to the first attenuation factor and the second attenuation factor;

[0138] Determine a data association relationship according to the third change attribute and a fourth change attribute corresponding to the historical date to obtain a second similarity;

[0139] The first similarity and the second similarity are integrated to determine the target associated date corresponding to the target predicted date.

[0140] In some implementations, the gas station data integration device 200 based on big data may be applied to a terminal device.

[0141] It should be noted that technical personnel in the relevant field can clearly understand that, for the convenience and conciseness of description, the specific working process of the gas station data integration device 200 based on big data described above can refer to the corresponding process in the aforementioned gas station data integration method embodiment based on big data, and will not be repeated here.

[0142] 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.

[0143] 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.

[0144] 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.

[0145] 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.

[0146] 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.

[0147] The processor is used to run a computer program stored in the memory, and implement any one of the gas station data integration methods based on big data provided by the embodiments of the present invention when executing the computer program.

[0148] 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:

[0149] Obtaining first attribute data and first feature data corresponding to a target gas station on a historical date, wherein the first feature data is used to characterize a balance index corresponding to the target gas station, and the balance index is used to characterize the number of items sold at the target gas station;

[0150] Determine second attribute data related to item configuration of the target gas station according to the first attribute data and the first characteristic data;

[0151] Performing information alignment on the historical date according to the second attribute data to obtain a close date corresponding to the historical date;

[0152] Obtain the third attribute data and the second characteristic data corresponding to the target gas station on the similar date;

[0153] The first feature data and the second feature data are combined to obtain third feature data, and a difference calculation is performed on any two first sub-data in the third feature data to obtain a target difference;

[0154] Performing data adjustment on the third characteristic data according to the target difference to obtain fourth characteristic data;

[0155] Obtaining associated sub-data corresponding to any second sub-data in the fourth characteristic data within a target range, and performing abnormal screening on the third characteristic data according to the associated sub-data to obtain target characteristic data corresponding to the target gas station under the historical date and the similar date;

[0156] Performing abnormality identification on the second attribute data and the third attribute data to obtain target attribute data corresponding to the historical date and the similar date;

[0157] Data integration is performed according to the target feature data and the target attribute data to determine a data integration result corresponding to the target gas station.

[0158] In some implementations, during the process of determining the second attribute data related to the object configuration of the target gas station according to the first attribute data and the first feature data, the processor 301 executes:

[0159] Obtaining first related data corresponding to the first attribute data under a first data type and second related data corresponding to a second data type, wherein the first data type is discrete data and the second data type is categorical data;

[0160] calculating a first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data;

[0161] calculating a second degree of association between the third sub-data and the first characteristic data, and determining a third degree of association between the fourth sub-data and the first characteristic data according to the first degree of association and the second degree of association;

[0162] The first attribute data is screened according to the second degree of association and the third degree of association to obtain the second attribute data related to the target gas station and the item configuration.

[0163] In some implementations, during the process of calculating the first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data, the processor 301 executes:

[0164] Calculate the distance information between any two of the third sub-data in the first related data to obtain a target distance, and determine a first matrix according to the target distance;

[0165] Calculating the degree of difference between any two fourth sub-data in the second related data to obtain a target difference value, and determining a second matrix according to the first matrix and the target difference value;

[0166] Performing data fusion according to the first matrix and the second matrix to obtain a target distance between the third sub-data and the fourth sub-data;

[0167] The first degree of association between the second related data and the first related data is determined according to the target distance.

[0168] In some implementations, during the process of aligning the historical date according to the second attribute data to obtain a similar date corresponding to the historical date, the processor 301 executes:

[0169] Obtaining a first inherent attribute corresponding to the historical date from the second attribute data, and determining a first entity representation corresponding to the historical date according to the first inherent attribute;

[0170] Obtaining a second inherent attribute corresponding to the date to be determined other than the historical date from the second attribute data, and determining a second entity representation corresponding to the date to be determined according to the second inherent attribute;

[0171] Obtaining a first change attribute corresponding to the historical date from the second attribute data, and performing entity enhancement on the first entity representation according to the first change attribute to obtain a third entity representation;

[0172] Obtaining a second change attribute corresponding to the to-be-determined date from the second attribute data, and performing entity enhancement on the second entity representation according to the second change attribute to obtain a fourth entity representation;

[0173] Determine the date similarity between the historical date and the date to be determined according to the third entity representation and the fourth entity representation;

[0174] The close date corresponding to the historical date is selected from the dates to be determined according to the date similarity.

[0175] In some implementations, during the process of performing entity enhancement on the first entity representation according to the first change attribute to obtain the third entity representation, the processor 301 executes:

[0176] Obtaining first enhanced information obtained by enhancing the first entity representation in the entity relationship using a relationship-aware network of an entity enhancement model;

[0177] Calculating second enhancement information between the first changed attribute and the first entity representation using an attribute-aware network of the entity enhancement model;

[0178] Using the information merging network of the entity enhancement model, the first enhancement information and the second enhancement information are merged to obtain target enhancement information;

[0179] The information enhancement attention network of the entity enhancement model is used to perform entity enhancement on the first entity representation according to the target enhancement information to obtain the third entity representation.

[0180] In some implementations, the processor 301, in the process of performing abnormal screening on the third characteristic data according to the associated sub-data to obtain the target characteristic data corresponding to the target gas station on the historical date and the similar date, executes:

[0181] Performing information entropy calculation on the associated sub-data to obtain entropy value information, and calculating variance information corresponding to the associated sub-data;

[0182] Determine a target representation value corresponding to the second sub-data within the target range according to the entropy value information and the variance information;

[0183] Performing abnormal screening on the third characteristic data according to the target characterization value to obtain the target characteristic data corresponding to the target gas station on the historical date and the similar date;

[0184] The target characterization value is obtained according to the following formula:

[0185]

[0186] in, represents the weight corresponding to the entropy value information, represents the weight corresponding to the variance information, m represents the target range, represents the frequency information corresponding to the j-th second sub-data, represents the data corresponding to the j-th second sub-data, represents the data corresponding to the t-th second sub-data, and e represents a natural constant.

[0187] In some implementations, after obtaining the data integration result, the processor 301 further executes:

[0188] Determine a target prediction date corresponding to the target user, and obtain a third inherent attribute and a third change attribute corresponding to the target prediction date;

[0189] Obtaining a target associated date corresponding to the target forecast date from the historical dates according to the third inherent attribute and the third variable attribute;

[0190] Obtaining prediction feature data corresponding to the target prediction date from the data integration result according to the target association date;

[0191] A resource scheduling strategy corresponding to the target forecast date is determined according to the forecast characteristic data.

[0192] In some implementations, during the process of obtaining the target associated date corresponding to the target predicted date from the historical date according to the third inherent attribute and the third variable attribute, the processor 301 executes:

[0193] Determining a first attenuation factor and a second attenuation factor, and determining a first similarity between the target forecast date and the historical date using the third inherent attribute according to the first attenuation factor and the second attenuation factor;

[0194] Determine a data association relationship according to the third change attribute and a fourth change attribute corresponding to the historical date to obtain a second similarity;

[0195] The first similarity and the second similarity are integrated to determine the target associated date corresponding to the target predicted date.

[0196] It should be noted that technical personnel in the relevant field can clearly understand that, for the convenience and conciseness of description, the specific working process of the terminal device described above can refer to the corresponding process in the aforementioned gas station data integration method embodiment based on big data, and will not be repeated here.

[0197] 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 any step of the gas station data integration method based on big data provided in the description of the embodiment of the present invention.

[0198] The storage medium may be an internal storage unit of the terminal device in the aforementioned 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 (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal device.

[0199] 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.

[0200] 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.

[0201] 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 are only specific embodiments 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 gas station data integration method based on big data, characterized in that: The method comprises: Obtaining first attribute data and first feature data corresponding to a target gas station on a historical date, wherein the first feature data is used to characterize a balance index corresponding to the target gas station, and the balance index is used to characterize the number of items sold at the target gas station; Determine second attribute data related to item configuration of the target gas station according to the first attribute data and the first characteristic data; Performing information alignment on the historical date according to the second attribute data to obtain a close date corresponding to the historical date; Obtain the third attribute data and the second characteristic data corresponding to the target gas station on the similar date; The first feature data and the second feature data are combined to obtain third feature data, and a difference calculation is performed on any two first sub-data in the third feature data to obtain a target difference; Performing data adjustment on the third characteristic data according to the target difference to obtain fourth characteristic data; Obtaining associated sub-data corresponding to any second sub-data in the fourth characteristic data within a target range, and performing abnormal screening on the third characteristic data according to the associated sub-data to obtain target characteristic data corresponding to the target gas station under the historical date and the similar date; Performing abnormality identification on the second attribute data and the third attribute data to obtain target attribute data corresponding to the historical date and the similar date; Data integration is performed according to the target feature data and the target attribute data to determine a data integration result corresponding to the target gas station.

2. The method according to claim 1, characterized in that: The determining, according to the first attribute data and the first feature data, second attribute data related to the configuration of the object at the target gas station includes: Obtaining first related data corresponding to the first attribute data under a first data type and second related data corresponding to a second data type, wherein the first data type is discrete data and the second data type is categorical data; calculating a first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data; calculating a second degree of association between the third sub-data and the first characteristic data, and determining a third degree of association between the fourth sub-data and the first characteristic data according to the first degree of association and the second degree of association; The first attribute data is screened according to the second degree of association and the third degree of association to obtain the second attribute data related to the target gas station and the item configuration.

3. The method according to claim 2, characterized in that The calculating a first degree of association between any third sub-data in the first related data and any fourth sub-data in the second related data includes: Calculate the distance information between any two of the third sub-data in the first related data to obtain a target distance, and determine a first matrix according to the target distance; Calculating the degree of difference between any two fourth sub-data in the second related data to obtain a target difference value, and determining a second matrix according to the first matrix and the target difference value; Performing data fusion according to the first matrix and the second matrix to obtain a target distance between the third sub-data and the fourth sub-data; The first degree of association between the second related data and the first related data is determined according to the target distance.

4. The method according to claim 1, characterized in that: The step of performing information alignment on the historical date according to the second attribute data to obtain a close date corresponding to the historical date includes: Obtaining a first inherent attribute corresponding to the historical date from the second attribute data, and determining a first entity representation corresponding to the historical date according to the first inherent attribute; Obtaining a second inherent attribute corresponding to the date to be determined other than the historical date from the second attribute data, and determining a second entity representation corresponding to the date to be determined according to the second inherent attribute; Obtaining a first change attribute corresponding to the historical date from the second attribute data, and performing entity enhancement on the first entity representation according to the first change attribute to obtain a third entity representation; Obtaining a second change attribute corresponding to the to-be-determined date from the second attribute data, and performing entity enhancement on the second entity representation according to the second change attribute to obtain a fourth entity representation; Determine the date similarity between the historical date and the date to be determined according to the third entity representation and the fourth entity representation; The close date corresponding to the historical date is selected from the dates to be determined according to the date similarity.

5. The method according to claim 4, characterized in that The step of performing entity enhancement on the first entity representation according to the first change attribute to obtain the third entity representation includes: Obtaining first enhanced information obtained by enhancing the first entity representation in the entity relationship using a relationship-aware network of an entity enhancement model; Calculating second enhancement information between the first changed attribute and the first entity representation using an attribute-aware network of the entity enhancement model; Using the information merging network of the entity enhancement model, the first enhancement information and the second enhancement information are merged to obtain target enhancement information; The information enhancement attention network of the entity enhancement model is used to perform entity enhancement on the first entity representation according to the target enhancement information to obtain the third entity representation.

6. The method according to claim 1, characterized in that The performing abnormal screening on the third characteristic data according to the associated sub-data to obtain the target characteristic data corresponding to the target gas station on the historical date and the similar date includes: Performing information entropy calculation on the associated sub-data to obtain entropy value information, and calculating variance information corresponding to the associated sub-data; Determine a target representation value corresponding to the second sub-data within the target range according to the entropy value information and the variance information; Performing abnormal screening on the third characteristic data according to the target characterization value to obtain the target characteristic data corresponding to the target gas station on the historical date and the similar date; The target characterization value is obtained according to the following formula: ; in, represents the weight corresponding to the entropy value information, represents the weight corresponding to the variance information, m represents the target range, represents the frequency information corresponding to the j-th second sub-data, represents the data corresponding to the j-th second sub-data, represents the data corresponding to the t-th second sub-data, and e represents a natural constant.

7. The method according to claim 1, characterized in that After obtaining the data integration result, the method further includes: Determine a target prediction date corresponding to the target user, and obtain a third inherent attribute and a third change attribute corresponding to the target prediction date; Obtaining a target associated date corresponding to the target forecast date from the historical dates according to the third inherent attribute and the third variable attribute; Obtaining prediction feature data corresponding to the target prediction date from the data integration result according to the target association date; A resource scheduling strategy corresponding to the target forecast date is determined according to the forecast characteristic data.

8. The method according to claim 7, characterized in that The step of obtaining the target associated date corresponding to the target forecast date from the historical date according to the third inherent attribute and the third change attribute includes: Determining a first attenuation factor and a second attenuation factor, and determining a first similarity between the target forecast date and the historical date using the third inherent attribute according to the first attenuation factor and the second attenuation factor; Determine a data association relationship according to the third change attribute and a fourth change attribute corresponding to the historical date to obtain a second similarity; The first similarity and the second similarity are integrated to determine the target associated date corresponding to the target predicted date.

9. A gas station data integration device based on big data, characterized in that: include: A data acquisition module, used to obtain first attribute data and first feature data corresponding to a target gas station on a historical date, wherein the first feature data is used to characterize a balance index corresponding to the target gas station, and the balance index is used to characterize the number of items sold at the target gas station; A data screening module, used for determining second attribute data related to the configuration of items at the target gas station according to the first attribute data and the first characteristic data; A data alignment module, used for performing information alignment on the historical date according to the second attribute data to obtain a close date corresponding to the historical date; An information acquisition module, used to obtain the third attribute data and the second characteristic data corresponding to the target gas station on the similar date; A difference calculation module, used for combining the first feature data and the second feature data to obtain third feature data, and calculating the difference between any two first sub-data in the third feature data to obtain a target difference; A data adjustment module, used for adjusting the third characteristic data according to the target difference to obtain fourth characteristic data; an abnormality screening module, used to obtain the associated sub-data corresponding to any second sub-data in the fourth characteristic data within the target range, and perform abnormality screening on the third characteristic data according to the associated sub-data, to obtain the target characteristic data corresponding to the target gas station under the historical date and the similar date; An abnormality identification module, used for performing abnormality identification on the second attribute data and the third attribute data, and obtaining target attribute data corresponding to the historical date and the similar date; The data integration module is used to perform data integration according to the target feature data and the target attribute data, and determine the data integration result corresponding to the target gas station.

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 gas station data integration method based on big data as described in any one of claims 1 to 8 when executing the computer program.

Images