A website visit volume prediction method, device and equipment

By constructing the change relationship between website visits and extracting characteristics, and determining data categories, accurate prediction of website visits is achieved, the problem of resource waste and paralysis risks in the existing technology is solved, and the reasonable allocation of resources and prediction accuracy is ensured.

CN113158117BActive Publication Date: 2024-11-12INDUSTRIAL AND COMMERCIAL BANK OF CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110472701.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-04-29
Publication Date
2024-11-12
Estimated Expiration
2041-04-29

AI Technical Summary

Technical Problem

When predicting website visits, the existing technology directly conducts incremental predictions based on real-time data, resulting in waste of resources and the inability to effectively deal with sudden increase in visits, which poses a risk of website paralysis and lacks accurate prediction methods.

Method used

By obtaining historical access data, constructing a change relationship of visits, extracting visit characteristics, determining data categories, using characteristics and categories to predict expected visits, and improving prediction accuracy.

Benefits of technology

It improves the accuracy of website visits prediction, ensures the reasonable allocation of network and computing resources, and effectively avoids the paralysis of the website under the impact of large traffic.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113158117B_ABST
    Figure CN113158117B_ABST
Patent Text Reader

Abstract

The embodiments of this specification provide a website visit volume prediction method, device and equipment, which can be applied to the field of big data technology. The method includes: obtaining historical visit data; the historical visit data includes the visit volume corresponding to the target website at at least two moments in the sampling period; constructing a visit volume change relationship based on the historical visit data; the visit volume change relationship is used to represent the change of the visit volume over time; extracting at least one visit volume feature from the visit volume change relationship; the visit volume feature is used to describe the change trend of the visit volume; using the visit volume feature to determine the data category corresponding to the historical visit data; predicting the expected visit volume of the target website through the data category and the visit volume feature. The above method predicts the visit volume according to the change trend reflected by the category and the visit volume situation, thereby improving the accuracy of the visit volume prediction and ensuring the reasonable allocation of network resources and computing resources.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present specification relate to the field of big data technology, and in particular to a website visit volume testing method, device and equipment. Background Art

[0002] With the development of computer technology and information technology, the Internet has become an important way for users to obtain services. Various types of websites have emerged to meet the different needs of users. Different websites also have different traffic. Some popular websites often have large traffic and have high requirements for network bandwidth and servers. Therefore, it is of great significance to configure the corresponding network and computing equipment by predicting the website traffic, which can effectively avoid the website from being paralyzed under the impact of large traffic.

[0003] At present, when predicting website visits, the incremental prediction results are often used as the final prediction results directly based on the website visit data obtained in real time. Although this prediction method can guarantee the buffering capacity of the website to a certain extent, it often requires additional configuration of limited computing resources and network bandwidth, resulting in a waste of resources. In addition, it is impossible to effectively deal with the sudden increase in visit volume, and there is also a certain risk of website paralysis. Therefore, there is an urgent need for a method that can accurately predict website visits. Summary of the invention

[0004] The purpose of the embodiments of this specification is to provide a method, device and equipment for predicting website visit volume, so as to solve the problem of how to improve the prediction accuracy of website visit volume.

[0005] To solve the above technical problems, an embodiment of the present specification provides a website visit volume prediction method, comprising: obtaining historical visit data; the historical visit data includes the visit volume corresponding to the target website at at least two moments in a sampling period; constructing a visit volume change relationship based on the historical visit data; the visit volume change relationship is used to represent the change of the visit volume over time; extracting at least one visit volume feature from the visit volume change relationship; the visit volume feature is used to describe the change trend of the visit volume; using the visit volume feature to determine the data category corresponding to the historical visit data; and predicting the expected visit volume of the target website through the data category and the visit volume feature.

[0006] The embodiment of this specification also proposes a website visit volume prediction device, including: a historical visit data acquisition module, used to acquire historical visit data; the historical visit data includes the visit volume corresponding to the target website at at least two moments in a sampling period; a visit volume change relationship construction module, used to construct a visit volume change relationship based on the historical visit data; the visit volume change relationship is used to represent the change of the visit volume over time; a visit volume feature extraction module, used to extract at least one visit volume feature from the visit volume change relationship; the visit volume feature is used to describe the change trend of the visit volume; a data category determination module, used to determine the data category corresponding to the historical visit data using the visit volume feature; an expected visit volume prediction module, used to predict the expected visit volume of the target website through the data category and the visit volume feature.

[0007] The embodiments of this specification also propose a website visit volume prediction device, including a memory and a processor; the memory is used to store computer program instructions; the processor is used to implement the following steps based on the computer program instructions: obtaining historical visit data; the historical visit data includes the visit volume corresponding to the target website at at least two moments in a sampling period; constructing a visit volume change relationship based on the historical visit data; the visit volume change relationship is used to represent the change of the visit volume over time; extracting at least one visit volume feature from the visit volume change relationship; the visit volume feature is used to describe the change trend of the visit volume; using the visit volume feature to determine the data category corresponding to the historical visit data; predicting the expected visit volume of the target website through the data category and the visit volume feature.

[0008] It can be seen from the technical solutions provided by the above embodiments of this specification that the embodiments of this specification construct corresponding visit volume change relationships based on the historical visit data of the target website, and classify different historical visit data according to the visit volume characteristics reflected in the visit volume change relationship, and the prediction of the visit volume corresponding to the target website can be effectively realized through the data categories and visit volume characteristics. The above method can intuitively obtain the categories corresponding to the historical visit data in the change graph, and then realize the prediction of the visit volume according to the change trend reflected by the category and the visit volume situation, thereby improving the accuracy of the visit volume prediction and ensuring the reasonable allocation of network resources and computing resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0010] Figure 1 This is a flow chart of a website visit volume prediction method according to an embodiment of this specification;

[0011] Figure 2A This is a schematic diagram of the rising section of a visit volume change relationship according to an embodiment of this specification;

[0012] Figure 2B This is a schematic diagram of a descending section of a visit volume change relationship according to an embodiment of this specification;

[0013] Figure 3A A line graph of access data according to an embodiment of this specification;

[0014] Figure 3B A bar chart of access data according to an embodiment of this specification;

[0015] Figure 3C A schematic diagram of a pageview change relationship according to an embodiment of this specification;

[0016] Figure 4A This is a visit volume change graph based on the visit volume at the starting point according to the embodiment of this specification;

[0017] Figure 4B This is a visit volume change graph based on the maximum visit volume according to the embodiment of this specification;

[0018] Figure 4C This is a visit volume change graph based on the minimum visit volume according to an embodiment of this specification;

[0019] Figure 4D This is a visit volume change graph based on the visit volume of the termination point according to the embodiment of this specification;

[0020] Figure 5 A schematic diagram of a data set according to an embodiment of this specification;

[0021] Figure 6 A schematic diagram of a prediction result of an embodiment of this specification;

[0022] Figure 7 This is a flowchart of a website visit volume prediction process according to an embodiment of this specification;

[0023] Figure 8 This is a module diagram of a website visit volume prediction device according to an embodiment of this specification;

[0024] Fig. 9 This is a structural diagram of a website visit volume prediction device according to an embodiment of this specification. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of this specification to clearly and completely describe the technical solutions in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0026] In order to solve the above technical problems, the embodiments of this specification propose a website visit volume prediction method. The execution subject of the website visit volume prediction method can be a website visit volume prediction device, which includes but is not limited to a server, an industrial computer, a PC, etc. Figure 1 As shown, the website visit volume prediction method may include the following specific implementation steps.

[0027] S110: Acquire historical access data; the historical access data includes the number of visits corresponding to the target website at at least two moments in a sampling period.

[0028] The historical access data may be data corresponding to the historical access situation of the target website. Specifically, the historical access data may be, for example, the access volume data of the target website at various times.

[0029] The target website may be a website for which visit volume prediction is required, and the target website may include only a single web page or may include multiple web pages, without limitation.

[0030] In some implementations, a certain sampling period may be set when extracting the historical access data. That is, the access data for the target website within the sampling period is used as the historical access data. The sampling period may be, for example, based on the n days before the current moment, or may be a pre-set fixed time period, and there is no limitation on this.

[0031] In some implementations, a corresponding sampling moment may be set within the sampling period. The sampling moment may be a sampling moment set based on a fixed time interval within the sampling period, such as setting a sampling moment every one minute to collect historical access data, or sampling moments with different time intervals may be set based on different time periods, such as setting a shorter sampling moment during the day and a longer sampling moment at night to obtain a better sampling effect. In practical applications, the sampling moment may be set according to demand, and is not limited to the above examples, which will not be described in detail here.

[0032] S120: constructing a visit volume change relationship based on the historical visit data; the visit volume change relationship is used to represent the change of the visit volume over time.

[0033] After obtaining the historical access data, a visit volume change relationship can be constructed based on the historical access data. The visit volume change relationship can be used to reflect the change of the visit volume of the target website over time. Based on the visit volume change relationship, certain features can be extracted to achieve analysis of the visit volume of the target website.

[0034] In some embodiments, the visit volume change relationship may be a visit volume change relationship based on a candlestick chart. A candlestick chart is a schematic diagram used to describe price changes in the stock market or futures market. The candlestick chart of stocks has four important factors: opening price, highest price, lowest price, and closing price, which reflect changes in the general trend. Correspondingly, the visit volume change relationship may also include the visit volume at the start time, the visit volume at the end time, the maximum visit volume, and the minimum visit volume. Based on these characteristic values ​​in the visit volume change relationship, the characteristics corresponding to the visit volume change relationship can be effectively determined, thereby determining the category corresponding to the historical visit data. Figure 2A As shown in the figure, it is the eigenvalue size corresponding to the rising stage, such as Figure 2B As shown, the size of the eigenvalues ​​corresponding to the descending stage is shown, so that the data change situation can be determined based on the specific sizes of these four eigenvalues.

[0035] In some embodiments, the method for constructing the visit volume change relationship may be to first draw a vertical bar graph using the visit volume in the historical visit data, and then connect the columns in the vertical bar graph based on the horizontal visual graph algorithm to obtain the visit volume change relationship. Among them, the vertical bar graph may use multiple columns to respectively represent the specific values ​​of the visit volume at each moment. The visit volume change relationship is based on the columns in the vertical bar graph, by connecting each column and ensuring that the connecting line does not intersect with the column, to construct a corresponding connection graph as the final visit volume change relationship.

[0036] Let's use a specific example to illustrate. Figure 3A As shown in FIG. 1 , it is a graph of the changes in the number of visits corresponding to 12 different moments, and the size of the ordinate of each point can be used to reflect the specific size of the number of visits. Figure 3B As shown, Figure 3A The corresponding visit volume at each moment is converted into a bar chart for display. Figure 3C is Figure 3B Based on the bar graph, we first connect each adjacent column. After that, we can analyze whether different columns can be connected without intersecting the middle column, and repeat this process until all the connections are completed, and we get the following Figure 3C The connection effect diagram in .

[0037] Correspondingly, such as Figure 4AAs shown in , the visit volume change relationship is constructed based on the visit volume at the start time; Figure 4B As shown in , it is the visit volume change relationship based on the maximum visit volume; Figure 4C As shown in , it is the visit volume change relationship based on the minimum visit volume; Figure 4D As shown in FIG. 1 , the visit volume change relationship is constructed based on the visit volume at the end time. Based on these visit volume change relationships, the overall change of the data can be reflected.

[0038] S130: extracting at least one pageview feature from the pageview change relationship; the pageview feature is used to describe a change trend of the pageview.

[0039] After the visit volume variation relationship is drawn, the visit volume feature can be extracted from the visit volume variation graph. The visit volume feature is used to describe the variation trend of the visit volume, so the visit volume feature can be the feature value corresponding to the point with strong representativeness in the visit volume variation relationship.

[0040] In some embodiments, the visit volume feature may include at least one of a starting point, an end point, a maximum point, and a minimum point in a visit volume change relationship. The starting point may be a point corresponding to the moment when the sampling period starts, and the end point may be a point corresponding to the moment when the sampling period ends. The maximum point may be a point corresponding to the maximum visit volume in the visit volume change relationship, and the minimum point may be a point corresponding to the minimum visit volume in the visit volume change relationship. The maximum point and the minimum point may better reflect the change range of the visit volume within the sampling period. Based on the setting of the sampling period, the starting point and the end point can also effectively show the visit volume change characteristics of the sampling period.

[0041] Based on the example in step S120, when the visit volume change relationship is a K-line chart, the starting point, end point, maximum point and minimum point can be easily and quickly extracted from the visit volume change relationship, thereby further accelerating the speed of obtaining visit volume features and improving the execution efficiency of the prediction process.

[0042] A specific method of obtaining the visit volume feature may be to realize automatic identification of the visit volume change relationship based on a machine learning model. Specifically, feature points corresponding to the visit volume feature may be marked in a sample graph to train a corresponding model. When it is necessary to extract the visit volume feature, the visit volume change relationship may be directly input into the corresponding model to output the identified visit volume feature.

[0043] In some embodiments, the visit volume feature may further include a feature point clustering coefficient and a change graph clustering coefficient. The feature point clustering coefficient may represent the connection relationship between the point corresponding to the visit volume at each moment and other points. Since different feature points have different correlations in the visit volume change relationship, the feature point clustering coefficient may be obtained by clustering the relationship between these feature points. The change graph clustering coefficient may be the clustering coefficient obtained by integrating the feature point clustering coefficients of each point, and is used to reflect the characteristics of the entire visit volume change relationship as a whole.

[0044] Specifically, we can use the formula Calculate the clustering coefficient of the feature points, where Clu_coef(i) is the clustering coefficient of the feature point i, D i is the number of vertices connected to feature point i in the visit volume change relationship, and E is the number of edges connected in the visit volume change relationship after removing feature point i.

[0045] The parameters in the above formula are explained using a specific example in conjunction with the accompanying drawings. Figure 3C As shown, there are 11 vertices in the graph, and the maximum number of edges that can be formed by each vertex and other adjacent vertices is D i For example, there are three edges connected to vertex 1. The number of edges connected in the undirected graph after removing the vertex is E. For example, there are 12 edges after removing vertex 1. Based on the drawn undirected graph, the parameters for calculating the clustering coefficient of the feature points of each vertex can be obtained.

[0046] Accordingly, we can use the formula Where Clu_coef(G) is the clustering coefficient of the change graph, and N is the number of feature points. The above calculation method is to calculate the clustering coefficient of the feature points of each vertex, and then take the average value of the clustering coefficients of these feature points to obtain the clustering features corresponding to the entire change graph.

[0047] By obtaining the characteristic point clustering coefficient and / or the change graph clustering coefficient, the connections between the points in the visit volume change relationship can be effectively summarized, so that the characteristics reflected in the visit volume change relationship can be effectively utilized to achieve accurate prediction of the visit volume.

[0048] It should be noted that the visit volume features obtained in the above two implementations can be combined during the application process. For example, after extracting the starting point, end point, maximum point and minimum point in the visit volume change relationship, only the feature point clustering coefficient is calculated for these points, and the change graph clustering coefficient is obtained comprehensively, thereby improving the processing efficiency while ensuring the processing effect.

[0049] In actual applications, other types of pageview features may be selected as needed, and are not limited to the above examples, which will not be described in detail here.

[0050] S140: Determine the data category corresponding to the historical access data using the access volume feature.

[0051] Since the historical access data may contain access data corresponding to multiple moments, the access data may have complex changes during the sampling period. It may be difficult to classify the data directly based on the historical access data. Therefore, after obtaining the access features, since the access features themselves can accurately and intuitively reflect the changes in access data during this sampling period, using the access features to determine the data category corresponding to the historical access data not only ensures the accuracy of the classification to a certain extent, but also simplifies the difficulty of classifying the historical access data.

[0052] The data category can be a category based on the number of visits and the trend of the number of visits. The data category can be used to define different categories of historical access data. Accordingly, different data categories can reflect different trends of the number of visits, so that accurate prediction of the number of visits can be achieved based on the number of visits in the sampling period.

[0053] The data categories can be set autonomously based on the experience of managers before executing a specific prediction process, or they can be categories obtained by classifying different sample data. For example, unsupervised learning can be performed on some sample data, and after these sample data are divided into different categories, managers can summarize these categories. The number of data categories can also be set based on test requirements or the number of layers of unsupervised learning. In actual applications, other methods can also be used to obtain the data categories, which are not limited to the above examples and will not be repeated here.

[0054] When determining the data category, the access volume feature may be input into a classification model to determine the data category corresponding to the historical access data. The classification model may be a pre-trained model for classifying access volume change relationships.

[0055] In some embodiments, if the corresponding data categories are predetermined, training sample data can be obtained first when training the classification model. The training sample data can be a visit volume change relationship obtained based on the website visit volume. These training sample data can correspond to data labels for indicating the categories corresponding to the training sample data. During the training process, multiple binary classifiers are trained based on these training sample data and data labels. The number of these binary classifications and the categories divided can correspond to the pre-set data categories. In the training stage, the training of binary classifiers of different categories can be implemented in turn. Correspondingly, different binary classifiers can also be used in turn in the recognition stage to finally complete the recognition of the data categories of the historical access data.

[0056] Let's use a specific example to illustrate. Suppose we use support vector machine (SVM), a generalized classifier that uses supervised learning to classify data into two categories, to determine the data category. The idea is to find an optimal hyperplane that maximizes the distance between the two categories. The specific process can be to first use the visit volume change relationship to map the K-line model of a period of time or any type of K-line model in a certain period of time into an undirected graph, and then extract the clustering coefficient characteristics of the four elements of the K-line model as the training sample set. Let D = {(x k ,y k )|k=1,2,3,4,…,M,x k ∈R n ,y k ∈R}, where x k Represents the clustering coefficient characteristics of the four elements of a K-line chart, y k It indicates the type label of the K-line model. k The value of is 1 or -1, so that the samples marked as 1 and -1 are located on both sides of the decision plane. For the sample data, we only need to determine which side of the decision plane it is on to determine the classification label corresponding to the sample data. If it is located on the -1 layer of the decision plane, continue classification.

[0057] Secondly, for the selection of support vector machine parameters, commonly used kernel functions include linear kernel function, polynomial kernel function, S-type kernel function and Gaussian radial kernel function (RBF). In this example, the Gaussian kernel function is used to map the samples from n dimensions to n+1 or higher dimensions, so that the originally linearly inseparable data becomes linearly separable. The expression of the Gaussian kernel function is Among them, ‖xx c ‖ 2 It can be regarded as the Euclidean distance between two vectors, x c is the test point (i.e., the core value of a certain feature), and σ is the wide variable of the kernel bandwidth.

[0058] Correspondingly, assuming that k binary classifiers are trained to distinguish k data categories, the k binary classifiers can be used to classify the sample data in turn until the corresponding data category is obtained. If none of these binary classifiers complete the recognition, it is determined that the data does not belong to these categories and the recognition fails.

[0059] In some embodiments, after obtaining the classification model, the accuracy of the classification model can also be verified using a fold cross validation method. Fold cross validation is a statistical validation technique used in machine learning to evaluate the performance of machine learning models. It uses a subset of a data set to train it, and then uses a complementary subset of the data set that was not used for training to evaluate the performance of the model. It can ensure that the model correctly captures patterns from the data without considering interference from the data.

[0060] Let's use a specific example to illustrate. We use 100 websites as a data set and divide them into a test set and a training set. Figure 5 As shown in the figure. The training set is used to train the SVM model and test it on this model. Figure 6 As shown in the figure, the prediction results can be divided into four categories: TP, FP, FN, and TN. Among them, TP (True Positive): the prediction result is the positive class, which is actually the positive class; FP (False Positive): the prediction result is the positive class, but it is actually the negative class; FN (False negative): the prediction result is the negative class, but it is actually the positive class; TN (True negative): the prediction result is the negative class, which is actually the negative class.

[0061] In order to intuitively reflect the classification effect, three standard performance indicators are used to evaluate the classification performance, namely sensitivity, specificity and accuracy. Among them, sensitivity can be understood as how many of all positive classes are predicted as positive classes (positive class prediction is correct); specificity can be understood as how many of all negative classes are predicted as negative classes (negative class prediction is correct); accuracy can be understood as the number of correct classifications in all experiments. Using five-fold cross-validation, you can test 5 times, and then get the average value of the standard performance indicators. The larger the value, the more accurate the classification. Specifically, the calculation formulas for the above indicators are

[0062] Specifically, a certain verification threshold can be set. When the calculated value of the above-mentioned standard performance indicator is greater than the verification threshold, it indicates that the classification model has a certain accuracy and can be put into practical application. If it is not reached, it indicates that the accuracy of the current classification model is insufficient, and it is necessary to continue iterative training or redesign the classification model.

[0063] By verifying the classification model, the accuracy of the classification model is guaranteed, and the accuracy of the prediction results can be improved accordingly.

[0064] S150: Predicting the expected traffic of the target website based on the data category and the traffic characteristics.

[0065] After obtaining the data category, the expected visits to the target website can be predicted based on the visit volume characteristics of the historical visit data. In the case where the data category can be used to represent different visit volume change trends, effective prediction of the visit volume can be achieved based on the data category and visit volume characteristics.

[0066] Using a specific example, assuming that the data type corresponding to the historical access data is determined to be stable based on the access volume characteristics, the expected access volume of the target website can be predicted to remain unchanged based on the access volume corresponding to the access volume characteristics. If the data type is fast-growing, the access volume corresponding to the access volume characteristics can be increased by a certain amount as the predicted access volume.

[0067] The expected visits may be visits for a shorter period of time, for example, the data of the previous n days is used as historical visit data to predict the visits for the next day, that is, the visits for the n+1 day. The expected visits may also be used to predict the visits for the next few days. The specific prediction period can be set according to the needs of the actual application, and there is no limitation on this.

[0068] The above method is explained using a scenario example. Figure 7 As shown in FIG. 1 , after the K-line model is constructed using historical data, an undirected graph structure can be constructed using a horizontal visual graph algorithm. From the undirected graph, the corresponding features are extracted by calculating the clustering coefficient. Afterwards, the extracted features are classified using SVM to identify the K-line pattern corresponding to the historical data, so that the K-line pattern can be used to accurately predict the number of visits.

[0069] Through the description of the above embodiments and scenario examples, it can be seen that the method constructs a corresponding visit volume change relationship based on the historical visit data of the target website, and realizes the classification of different historical visit data according to the visit volume characteristics reflected in the visit volume change relationship, and the prediction of the visit volume corresponding to the target website can be effectively realized through the data category and visit volume characteristics. The above method can intuitively obtain the category corresponding to the historical visit data in the change graph, and then realize the prediction of the visit volume according to the change trend reflected by the category and the visit volume situation, thereby improving the accuracy of the visit volume prediction and ensuring the reasonable allocation of network resources and computing resources.

[0070] based on Figure 1 The corresponding website visit volume prediction method introduces a website visit volume prediction device in the embodiment of this specification. The website visit volume prediction device can be set in a website visit volume prediction device. Figure 8 As shown, the website visit volume prediction device may include the following modules.

[0071] The historical access data acquisition module 810 is used to acquire historical access data; the historical access data includes the number of visits corresponding to the target website at at least two moments in a sampling period.

[0072] The pageview change relationship building module 820 is used to build a pageview change relationship based on the historical pageview data; the pageview change relationship is used to represent the change of pageviews over time.

[0073] The pageview feature extraction module 830 is used to extract at least one pageview feature from the pageview change relationship; the pageview feature is used to describe the change trend of the pageview.

[0074] The data category determination module 840 is used to determine the data category corresponding to the historical access data by using the access volume feature.

[0075] The expected traffic prediction module 850 is used to predict the expected traffic of the target website according to the data category and the traffic characteristics.

[0076] based on Figure 1 Corresponding to the website visit volume prediction method, the embodiment of this specification provides a website visit volume prediction device. Fig. 9 As shown, the website visit volume prediction device may include a memory and a processor.

[0077] In this embodiment, the memory may be implemented in any appropriate manner. For example, the memory may be a read-only memory, a mechanical hard disk, a solid-state hard disk, or a USB flash drive, etc. The memory may be used to store computer program instructions.

[0078] In this embodiment, the processor can be implemented in any appropriate manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, etc. The processor can execute the computer program instructions to implement the following steps: obtaining historical access data; the historical access data includes the number of visits corresponding to the target website at at least two moments in the sampling period; constructing a visit change relationship based on the historical access data; the visit change relationship is used to represent the change of the visit over time; extracting at least one visit feature from the visit change relationship; the visit feature is used to describe the change trend of the visit; using the visit feature to determine the data category corresponding to the historical access data; predicting the expected visit of the target website through the data category and the visit feature.

[0079] It should be noted that the website visit volume prediction method, device and equipment described in the embodiments of this specification can be applied to the field of big data technology, and can also be applied to any other field except the field of big data technology, without limitation.

[0080] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also be aware that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0081] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0082] It can be known from the above description of the implementation mode that the technicians in this field can clearly understand that the present specification can be implemented by means of software plus the necessary first hardware platform. Based on such an understanding, the technical solution of the present specification can be essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present specification or some parts of the embodiments.

[0083] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0084] This specification can be used in many first or special computer system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc.

[0085] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0086] Although the present specification is described through embodiments, those skilled in the art will appreciate that there are many modifications and changes to the present specification without departing from the spirit of the present specification, and it is intended that the appended claims include these modifications and changes without departing from the spirit of the present specification.

Claims

1. A website visit volume prediction method, characterized in that: include: Acquire historical access data; the historical access data includes the number of visits corresponding to the target website at at least two times within a sampling period; Building a visit volume change relationship based on the historical visit data; The visit volume change relationship is used to represent the change of the visit volume over time; the construction of the visit volume change relationship based on the historical visit data includes: drawing a vertical bar graph using the visit volume; the columns in the vertical bar graph correspond to the visit volume at each moment; based on the horizontal visual graph algorithm, connecting the columns in the vertical bar graph to obtain the visit volume change relationship; Extracting at least one visit volume feature from the visit volume change relationship; the visit volume feature is used to describe the change trend of the visit volume; the visit volume feature includes a feature point clustering coefficient and a change graph clustering coefficient; extracting at least one visit volume feature from the visit volume change relationship includes: obtaining a connection relationship between each feature point in the visit volume change relationship and the visit volume change relationship; obtaining a feature point clustering coefficient corresponding to the feature point based on the connection relationship; and obtaining a change graph clustering coefficient corresponding to the visit volume change relationship by combining the feature point clustering coefficients. Determine the data category corresponding to the historical access data by using the access volume feature; The expected traffic of the target website is predicted by the data category and the traffic characteristics.

2. The method according to claim 1, characterized in that The pageview change relationship includes a pageview change graph based on a K-line graph.

3. The method according to claim 1, characterized in that The extracting at least one pageview feature from the pageview change relationship includes: At least one visit volume feature is extracted based on the feature points in the visit volume change relationship; the feature points include at least one of a starting point, an end point, a maximum value point and a minimum value point in the visit volume change relationship.

4. The method according to claim 1, characterized in that The obtaining of a feature point clustering coefficient corresponding to the feature point based on the connection relationship includes: Using the formula Calculate the clustering coefficient of the feature points, where Clu_coef(i) is the clustering coefficient of the feature point i, D i is the number of vertices connected to feature point i in the visit volume change relationship, and E is the number of edges connected in the visit volume change relationship after removing feature point i; The step of synthesizing the clustering coefficients of the feature points to obtain the clustering coefficients of the change graph corresponding to the visit volume change relationship includes: Using the formula Where Clu_coef(G) is the clustering coefficient of the change graph, and N is the number of feature points.

5. The method according to claim 1, characterized in that The determining the data category corresponding to the historical access data by using the access volume feature includes: Based on the access volume feature, a classification model is used to determine the data category corresponding to the historical access data.

6. The method according to claim 5, characterized in that The classification model is obtained in the following way: Acquire training sample data; the training sample data includes a visit volume change relationship constructed based on the website visit volume; the training sample data corresponds to a data label; the data label is used to indicate the data category corresponding to the training sample data; Training at least one binary classifier based on the data labels and the training sample data; the binary classifier is used to identify the visit volume change relationship of the corresponding category; The at least one binary classifier is integrated to obtain a classification model.

7. The method according to claim 5, characterized in that Before determining the data category corresponding to the historical access data by using the classification model, the method further includes: The accuracy of the classification model was verified using the folded cross validation method.

8. A website visit volume prediction device, characterized in that: include: A historical access data acquisition module, used to acquire historical access data; the historical access data includes the number of visits corresponding to the target website at at least two moments in a sampling period; A pageview change relationship building module, used to build a pageview change relationship based on the historical pageview data; The visit volume change relationship is used to represent the change of the visit volume over time; the construction of the visit volume change relationship based on the historical visit data includes: drawing a vertical bar graph using the visit volume; the columns in the vertical bar graph correspond to the visit volume at each moment; based on the horizontal visual graph algorithm, connecting the columns in the vertical bar graph to obtain the visit volume change relationship; A visit volume feature extraction module is used to extract at least one visit volume feature from the visit volume change relationship; the visit volume feature is used to describe the change trend of the visit volume; the visit volume feature includes a feature point clustering coefficient and a change graph clustering coefficient; the extraction of at least one visit volume feature from the visit volume change relationship includes: obtaining a connection relationship between each feature point in the visit volume change relationship and the visit volume change relationship; obtaining a feature point clustering coefficient corresponding to the feature point based on the connection relationship; and obtaining a change graph clustering coefficient corresponding to the visit volume change relationship by combining the feature point clustering coefficients. A data category determination module, used to determine the data category corresponding to the historical access data by using the access volume feature; The expected traffic prediction module is used to predict the expected traffic of the target website according to the data category and the traffic characteristics.

9. A website visit volume prediction device, comprising a memory and a processor; The memory is used to store computer program instructions; The processor is used to implement the following steps based on the computer program instructions: obtaining historical access data; the historical access data includes the visits corresponding to the target website at at least two moments in a sampling period; and constructing a visit change relationship based on the historical access data; The visit volume change relationship is used to represent the change of visit volume over time; the construction of the visit volume change relationship based on the historical visit data includes: using the visit volume to draw a vertical bar graph; the bars in the vertical bar graph correspond to the visit volume at each moment; based on the horizontal visual graph algorithm, the bars in the vertical bar graph are connected to obtain the visit volume change relationship; extracting at least one visit volume feature from the visit volume change relationship; the visit volume feature is used to describe the change trend of the visit volume; the visit volume feature includes a feature point clustering coefficient and a change graph clustering coefficient; extracting at least one visit volume feature from the visit volume change relationship includes: obtaining the connection relationship between each feature point in the visit volume change relationship and the visit volume change relationship; based on the connection relationship, obtaining the feature point clustering coefficient corresponding to the feature point; comprehensively obtaining the change graph clustering coefficient corresponding to the visit volume change relationship by combining the feature point clustering coefficient; using the visit volume feature to determine the data category corresponding to the historical visit data; predicting the expected visit volume of the target website through the data category and the visit volume feature.