Markov chain-based browsing track analysis method, apparatus and device, and medium

Abstract

The invention relates to the field of intelligent decision making and financial science and technology, and discloses a browsing track analysis method and device based on a Markov chain, computer equipment and a storage medium, the method comprises the following steps: identifying index data of a historical user from a historical browsing record, recording browsing states of the historical user at different time points, extracting historical effective indexes from the index data; constructing an index chain structure of the historical effective indexes, identifying index states of browsing behaviors of historical users from the historical effective indexes, and constructing a state space of the browsing behaviors by using the index states; calculating a state transition probability between the index states in the state space, and constructing a state transition matrix of the browsing behavior through the state transition probability; calculating an index transition probability of the current effective index, and analyzing an index browsing track of the current user by using the index transition probability; and sending the interested indexes to the current user. According to the method, accurate track analysis can be carried out by utilizing the historical browsing data of the user.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent decision-making and financial technology, and in particular to a browsing trajectory analysis method, device, computer equipment and computer-readable storage medium based on Markov chain. Background Art

[0002] Browsing trajectory analysis based on Markov chain refers to generating the user's historical browsing trajectory into a Markov chain, and then predicting the probability of the user's browsing trajectory in the future time period based on the probability size corresponding to the Markov chain. According to the predicted probability size of the browsing trajectory in the future time period, it is determined which browsing trajectories can be used as the user's future browsing trajectory. Browsing trajectory analysis based on Markov chain can be applied to the financial technology industry. For example, in the insurance industry, the probability size of the user's next browsing record can be analyzed based on the user's historical browsing record of insurance products, so as to determine which records the user will browse next.

[0003] At present, with the rapid development of the Internet and e-commerce, users' browsing behavior on websites has become more and more complicated. Accurately predicting users' browsing tracks is of great significance for improving user experience and optimizing website structure. This is because by accurately predicting users' browsing tracks, personalized recommendations for browsing pages can be made for users, thereby improving users' browsing comfort.

[0004] In the prior art, there is a lack of an effective method to utilize the user's historical browsing data for accurate trajectory analysis. Summary of the invention

[0005] The present invention provides a browsing trajectory analysis method, device, computer equipment and computer-readable storage medium based on Markov chain, the main purpose of which is to use the user's historical browsing data to perform accurate trajectory analysis.

[0006] In a first aspect, the present invention provides a browsing trajectory analysis method based on a Markov chain, comprising:

[0007] Collect historical browsing records of historical users, identify index data of the historical users from the historical browsing records, record browsing status of the historical users at different time points, and extract historical valid indicators from the index data based on the browsing status;

[0008] Constructing an indicator chain structure of the historical valid indicators, identifying the indicator state of the historical user's browsing behavior from the historical valid indicators, and constructing the state space of the browsing behavior using the indicator state;

[0009] Based on the indicator chain structure, the state transition probability between each indicator state in the state space is calculated, and the state transition matrix of the browsing behavior is constructed through the state transition probability;

[0010] Collecting the current browsing history of the current user, obtaining the current valid indicator of the current user from the current browsing history, calculating the indicator transfer probability of the current valid indicator according to the state transfer matrix, and analyzing the indicator browsing trajectory of the current user by using the indicator transfer probability;

[0011] Based on the indicator browsing trajectory, the indicator of interest is sent to the current user, so as to complete the indicator browsing trajectory analysis of the current user through the indicator of interest, and obtain the indicator browsing trajectory analysis result of the current user.

[0012] In a second aspect, the present invention further provides a browsing trajectory analysis device based on a Markov chain, the device comprising:

[0013] An indicator extraction module is used to collect historical browsing records of historical users, identify indicator data of the historical users from the historical browsing records, record browsing states of the historical users at different time points, and extract historical valid indicators from the indicator data based on the browsing states;

[0014] A state construction module, used to construct an indicator chain structure of the historical effective indicators, identify the indicator state of the historical user's browsing behavior from the historical effective indicators, and use the indicator state to construct a state space of the browsing behavior;

[0015] A matrix construction module, used to calculate the state transition probability between each indicator state in the state space based on the indicator chain structure, and construct the state transition matrix of the browsing behavior through the state transition probability;

[0016] A trajectory analysis module, used to collect the current browsing history of the current user, obtain the current valid index of the current user from the current browsing history, calculate the index transfer probability of the current valid index according to the state transfer matrix, and analyze the index browsing trajectory of the current user using the index transfer probability;

[0017] The result obtaining module is used to send the indicators of interest to the current user based on the indicator browsing trajectory, so as to complete the indicator browsing trajectory analysis of the current user through the indicators of interest and obtain the indicator browsing trajectory analysis result of the current user.

[0018] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned browsing trajectory analysis method based on Markov chain when executing the computer program.

[0019] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned browsing trajectory analysis method based on Markov chain are implemented.

[0020] It can be seen that the embodiment of the present invention identifies the indicator data of the historical user from the historical browsing records to classify the cumbersome historical browsing records and reduce the category complexity of the historical browsing records. The embodiment of the present invention constructs an indicator chain structure of the historical effective indicators to model the user browsing behavior using the Markov chain, and regards each browsing indicator as a state. The user's transition between different states is accurately captured and modeled. Compared with traditional methods, the Markov chain model can more accurately describe the dynamic characteristics of user behavior. The embodiment of the present invention calculates the state transition probability between each indicator state in the state space based on the indicator chain structure, so as to dynamically calculate the transition probability between states by analyzing a large amount of user historical browsing data. This method can adapt to changes in user behavior and ensure the accuracy and real-time nature of the transition probability, thereby improving the accuracy of the prediction. Furthermore, the embodiment of the present invention constructs a state transition matrix of the browsing behavior through the state transition probability, so as to convert the state transition probability Organized into a matrix form, an efficient state transfer matrix construction method is provided, the state transfer matrix can quickly perform matrix operations, thereby realizing efficient prediction of the user's future browsing trajectory. This efficient matrix construction method has significant advantages in processing large-scale user data. Further, the embodiment of the present invention analyzes the index browsing trajectory of the current user by using the index transfer probability, so as to not only predict the user's next behavior, but also perform multi-step prediction. By repeatedly applying the state transfer matrix, the user's future multi-step browsing trajectory can be predicted, providing a deeper behavioral insight for the recommendation system. The embodiment of the present invention sends the current user an interesting indicator based on the index browsing trajectory, so as to be used for the trajectory prediction result based on the Markov chain, and the prediction result is seamlessly integrated into the recommendation system, so as to recommend the user the indicator data or page of the relevant industry that may be of interest in real time. Compared with the static recommendation system, the real-time recommendation system of the present invention can provide more personalized and dynamic recommendation services according to the user's current browsing state and predicted future trajectory. Therefore, the browsing trajectory analysis method, device, computer equipment and computer-readable storage medium based on the Markov chain proposed in the embodiment of the present invention can use the user's historical browsing data for accurate trajectory analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of an application environment of a browsing trajectory analysis method based on a Markov chain provided in an embodiment of the present invention;

[0022] Figure 2 A schematic diagram of a flow chart of a browsing trajectory analysis method based on a Markov chain provided by an embodiment of the present invention;

[0023] Figure 3 A schematic diagram of a module of a browsing trajectory analysis device based on a Markov chain provided by an embodiment of the present invention;

[0024] Figure 4 is a schematic diagram of a structure of a computer device in one embodiment of the present invention;

[0025] Figure 5 It is another structural schematic diagram of a computer device in one embodiment of the present invention.

[0026] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0027] 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.

[0028] The embodiment of the present invention provides a browsing trajectory analysis method based on Markov chain, which can be applied in Figure 1In an application environment, wherein the client communicates with the server through a network, the server can collect historical browsing records of historical users through the client, identify the indicator data of the historical users from the historical browsing records, record the browsing status of the historical users at different time points, and extract historical valid indicators from the indicator data based on the browsing status; construct an indicator chain structure of the historical valid indicators, identify the indicator status of the browsing behavior of the historical users from the historical valid indicators, and use the indicator status to construct the state space of the browsing behavior; based on the indicator chain structure, calculate the state transition probability between each indicator state in the state space, and construct the state transition matrix of the browsing behavior through the state transition probability; collect the current browsing records of the current user, obtain the current valid indicators of the current user from the current browsing records, calculate the indicator transition probability of the current valid indicators according to the state transition matrix, and use the indicator transition probability to analyze the indicator browsing trajectory of the current user; based on the indicator browsing trajectory, send the interested indicators to the current user, so as to complete the indicator browsing trajectory analysis of the current user through the interested indicators, and obtain the indicator browsing trajectory analysis result of the current user.

[0029] Reference Figure 2 FIG. 1 is a flow chart of a method for analyzing browsing trajectories based on a Markov chain according to an embodiment of the present invention. In an embodiment of the present invention, the method for analyzing browsing trajectories based on a Markov chain comprises the following steps S1-S5:

[0030] S1. Collect historical browsing records of historical users, identify indicator data of the historical users from the historical browsing records, record browsing status of the historical users at different time points, and extract historical valid indicators from the indicator data based on the browsing status.

[0031] In an embodiment of the present invention, the historical user refers to the historical user who visited the browsed page, and the historical browsing record refers to the access log left by the historical user when browsing the browsed page. The historical user and the historical browsing record are data under specific application scenarios. If the selected application scenario is different, the historical user and the historical browsing record will also be different. For example, in the application scenario of financial technology, the historical user may be an insurance user who browses financial software, and the historical browsing record refers to the insurance-related page data browsed by the insurance user in the historical period.

[0032] Furthermore, the embodiment of the present invention identifies the index data of the historical user from the historical browsing records to classify the complicated historical browsing records, thereby reducing the complexity of the categories of the historical browsing records.

[0033] The indicator data refers to multiple categories of financial historical browsing records obtained by classifying the historical browsing records.

[0034] In one embodiment of the present invention, identifying the indicator data of the historical user from the historical browsing records includes: dividing the record indicators of the historical browsing records; obtaining the timestamps of the historical browsing records; sorting the historical browsing records based on the timestamps to obtain a browsing record sequence; identifying target browsing records belonging to the same record indicator in the browsing record sequence; and merging the target browsing records into indicator data when the timestamps of the target browsing records are continuous timestamps.

[0035] Among them, the record index refers to the basis for classifying the historical browsing records, such as COR, CER, and CLR of a certain type of insurance; the timestamp refers to the moment when the historical browsing record is recorded; the browsing record sequence refers to the order of the historical browsing records arranged in ascending order of the timestamps; and the continuous timestamp refers to a timestamp that is within a continuous time period and is not interrupted.

[0036] Optionally, the process of merging the target browsing records into indicator data is, for example: when the target browsing records all belong to the COR category of a certain insurance type, the total category corresponding to the target browsing records, that is, the COR category of a certain insurance type, is used as the indicator data.

[0037] Furthermore, in an embodiment of the present invention, the browsing status includes online, invisible, offline, on-hook, etc. when browsing financial pages.

[0038] In one embodiment of the present invention, extracting historical valid indicators from the indicator data based on the browsing state includes: querying the target browsing state corresponding to the indicator data in the browsing state; and when the target browsing state is a preset browsing state, using the indicator data as a historical valid indicator.

[0039] The target browsing state refers to the browsing state at the same time as the time when the indicator data appears, and the preset browsing state refers to the state in which the user performs effective page browsing, including online, invisible, etc.

[0040] Optionally, the process of querying the target browsing state corresponding to the indicator data in the browsing state is determined by the indicator data and the browsing state having the same timestamp.

[0041] S2. Construct an indicator chain structure of the historical valid indicators, identify the indicator status of the historical user's browsing behavior from the historical valid indicators, and construct a state space of the browsing behavior using the indicator status.

[0042] The embodiment of the present invention constructs an indicator chain structure of the historical effective indicators to model the user browsing behavior using a Markov chain, regards each browsing indicator as a state, and accurately captures and models the user's transitions between different states. Compared with traditional methods, the Markov chain model can more accurately describe the dynamic characteristics of user behavior.

[0043] In one embodiment of the present invention, constructing the indicator chain structure of the historical valid indicators includes: arranging the historical valid indicators in order of their timestamps to obtain a valid indicator sequence; and using the valid indicator sequence as the indicator chain structure.

[0044] Optionally, the process of arranging the historical valid indicators in order of their timestamps to obtain a valid indicator sequence refers to: arranging the historical valid indicators in order of their timestamps from small to large to obtain a sequence of historical valid indicators.

[0045] Furthermore, in an embodiment of the present invention, the indicator status refers to the indicator category of the historically effective indicator, including the COR, CER, CLR and other statuses of a certain type of insurance in the field of financial technology.

[0046] Optionally, constructing the state space of the browsing behavior using the indicator state includes: assuming that the state space is {A, B, C}, this state space refers to the set of all possible states defined in the Markov chain model, A represents COR, B represents CER, and C represents CLR.

[0047] S3. Based on the indicator chain structure, calculate the state transition probability between each indicator state in the state space, and construct the state transition matrix of the browsing behavior through the state transition probability.

[0048] The embodiment of the present invention calculates the state transition probability between each indicator state in the state space based on the indicator chain structure, so as to dynamically calculate the transition probability between states by analyzing a large amount of user historical browsing data. This method can adapt to changes in user behavior, ensure the accuracy and real-time nature of the transition probability, and thus improve the accuracy of the prediction.

[0049] In one embodiment of the present invention, the state transition probability between each indicator state in the state space is calculated based on the indicator chain structure, including: querying the number of paths between every two historical valid indicators and the total number of departures of each historical valid indicator from the indicator chain structure; and calculating the state transition probability between each indicator state in the state space according to the path number and the total number of departures using the following formula:

[0050]

[0051] Wherein, P(A→B) represents the state transition probability from the A indicator state to the B indicator state in the state space, and x A→B Indicates the number of paths from the A indicator state to the B indicator state, x A→ Indicates the total number of departures of indicator state A.

[0052] Among them, the path number refers to the number of times two historical valid indicators with a chronological order appear in the indicator chain structure. For example, if the two historical valid indicators with a chronological order in the indicator chain structure are , and there are three paths from the indicator state to the indicator state, then 3 is taken as the path number, and the total number of departures refers to the number of all paths departing from the indicator state. For example, if the A indicator state is the direct predecessor of other indicator states in the indicator chain structure, the number of direct successors of the A indicator state is taken as the total number of departures of the indicator state.

[0053] Furthermore, an embodiment of the present invention constructs a state transition matrix of the browsing behavior through the state transition probability, and provides an efficient state transition matrix construction method by organizing the state transition probability into a matrix form. The state transition matrix can quickly perform matrix operations, thereby achieving efficient prediction of the user's future browsing trajectory. This efficient matrix construction method has significant advantages when processing large-scale user data.

[0054] In one embodiment of the present invention, constructing the state transition matrix of the browsing behavior through the state transition probabilities includes: obtaining first transition probabilities belonging to the same starting point in the state transition probabilities; querying second transition probabilities belonging to the same ending point in the state transition probabilities; using the first transition probability as a row value and the second transition probability as a column value; and constructing the state transition matrix of the browsing behavior through the row values ​​and the column values ​​using the following matrix:

[0055]

[0056] Among them, P represents the state transition matrix, P(A→A) represents the state transition probability from the A indicator state to the A indicator state, P(A→B) represents the state transition probability from the A indicator state to the B indicator state, P(A→C) represents the state transition probability from the A indicator state to the C indicator state, P(B→A) represents the state transition probability from the B indicator state to the A indicator state, P(B→B) represents the state transition probability from the B indicator state to the B indicator state, P(B→C) represents the state transition probability from the B indicator state to the C indicator state, P(C→A) represents the state transition probability from the C indicator state to the A indicator state, P(C→B) represents the state transition probability from the C indicator state to the B indicator state, and P(C→C) represents the state transition probability from the C indicator state to the C indicator state.

[0057] Among them, the starting points of the row values, such as P(A→A), P(A→B), P(A→C) in the state transfer matrix, P(A→A), P(A→B), P(A→C) are all A, and the ending points of the column values, such as P(A→A), P(B→A), P(C→A) in the state transfer matrix, P(A→A), P(B→A), P(C→A) are all A.

[0058] S4. Collect the current browsing history of the current user, obtain the current valid indicator of the current user from the current browsing history, calculate the indicator transfer probability of the current valid indicator according to the state transfer matrix, and use the indicator transfer probability to analyze the indicator browsing trajectory of the current user.

[0059] In one embodiment of the present invention, the calculating the indicator transition probability of the current valid indicator according to the state transition matrix includes: constructing an indicator transition model of the current valid indicator using the following formula:

[0060]

[0061] Among them, P((D1,D2,D3,...,D n )→E) represents the indicator transfer model of the current effective indicator, D1,D2,D3,...,D n Represents a sequence of currently valid indicators of length n, D1, D2, D3, ..., D n ∈{A,B,C}, E represents any indicator state in the state space, P(E) represents the probability of E appearing, P(D1|E)=P(E→D1) represents the state transition probability from the E indicator state to the D1 indicator state, P((D2,D3,...,D n )∣(E,D1))=P((E,D1)→(D2,D3,...,D n )) indicates that the E, D1 indicator status is changed to D2, D3, ..., Dn The state transition probability of the indicator state, P(D1) represents the probability of D1 appearing, P(D1→(D2,D3,...,D n )) indicates the state of D1 indicator to D2, D3, ..., D n The state transition probability of the indicator state;

[0062] Obtain an unknown transition probability in the indicator transfer model; query a target transition probability corresponding to the unknown transition probability from the state transfer matrix; substitute the target transition probability into the indicator transfer model to calculate the indicator transfer probability corresponding to the target transition probability through the indicator transfer model.

[0063] The indicator transfer model refers to the probability calculation model of the constructed Markov chain without calculating the numerical value, in which the unknown transfer probability is the state transfer probability. It should be noted that, like P((E,D1)→(D2,D3,...,D n ))、P(D1→(D2,D3,...,D n )) Such a many-to-many probability form P(many→many) needs to be converted into a one-to-one form P(one→one) through prior probability, posterior probability or Bayesian probability. The specific conversion principle is the same as the above Similarly, the many-to-many form is decomposed into a one-to-one form, which will not be further elaborated here.

[0064] Furthermore, the embodiment of the present invention utilizes the indicator transition probability to analyze the indicator browsing trajectory of the current user, so as to not only predict the user's next behavior, but also perform multi-step prediction. By repeatedly applying the state transition matrix, the user's future multi-step browsing trajectory can be predicted, providing a deeper behavioral insight for the recommendation system.

[0065] In one embodiment of the present invention, the use of the indicator transition probability to analyze the indicator browsing trajectory of the current user includes: setting a transition probability threshold of the current user; when the indicator transition probability is greater than the transition probability threshold, obtaining an indicator state sequence corresponding to the indicator transition probability; identifying the remaining indicator sequences except the current valid indicators from the indicator state sequence; and using the remaining indicator sequences as the indicator browsing trajectory.

[0066] Among them, the indicator state sequence refers to the set of indicator states when calculating the indicator transfer probability. It should be noted that the setting of the transfer probability threshold needs to be set according to the actual scenario, for example, it is set to 0.8. The purpose of setting the transfer probability threshold is to screen some indicator transfer probabilities with larger numerical values ​​and larger probability values. These indicator transfer probabilities can reflect the tendency of user browsing trajectory. The setting of the transfer probability threshold is not further elaborated here.

[0067] Exemplarily, the process of identifying the remaining indicator sequences other than the current valid indicator from the indicator state sequence is, for example: querying the position of the current valid indicator in the indicator state sequence, for example, the indicator state sequence is composed of a digital sequence of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and a linear chain structure is formed between 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and except for 1 and 10, each number has a unique predecessor and successor, and the position of the current valid indicator is the position of 1, 2, 3, 4, 5, then 6, 7, 8, 9, 10 is the remaining indicator sequence other than the current valid indicator in the indicator state sequence.

[0068] S5. Based on the indicator browsing trajectory, sending the indicator of interest to the current user, so as to complete the indicator browsing trajectory analysis of the current user through the indicator of interest, and obtain the indicator browsing trajectory analysis result of the current user.

[0069] The embodiment of the present invention sends the interested indicators to the current user based on the indicator browsing trajectory for use in the trajectory prediction result based on the Markov chain, and seamlessly integrates the prediction result into the recommendation system, so as to recommend indicator data or pages of related industries that may be of interest to the user in real time. Compared with the static recommendation system, the real-time recommendation system of the present invention can provide more personalized and dynamic recommendation services according to the user's current browsing status and predicted future trajectory.

[0070] In one embodiment of the present invention, sending the indicator of interest to the current user based on the indicator browsing trajectory includes: obtaining a first indicator page corresponding to the indicator browsing trajectory; obtaining a second indicator page corresponding to the current valid indicator; calculating the page similarity between the first indicator page and the second indicator page; when the page similarity is greater than a preset similarity, using the first indicator page as the page of interest for the current user; and sending the page of interest to the current user to complete the process of sending the indicator of interest to the current user.

[0071] Among them, the first indicator page refers to all front-end pages related to the indicator browsing trajectory, and the second indicator page refers to all pages involved in the current browsing record corresponding to the current valid indicator. The page similarity can be calculated by a similarity algorithm, and the preset similarity can be set to 0.9, or it can be set according to the actual scenario, which will not be further elaborated here.

[0072] Optionally, the process of calculating the page similarity between the first indicator page and the second indicator page refers to the process of calculating the similarity between the image data of the first indicator page and the image data of the second indicator page. The page similarity can be determined by calculating the similarity of pixels in the image data. It should be noted that since the types of the first indicator pages are complex and the number is large, if all of the first indicator pages are blindly recommended to the current user, the recommended content will be relatively redundant. Therefore, some pages that conform to the user's previous browsing habits are selected through the page similarity to recommend pages of interest to the user.

[0073] It can be seen that the embodiment of the present invention identifies the indicator data of the historical user from the historical browsing records to classify the cumbersome historical browsing records and reduce the category complexity of the historical browsing records. The embodiment of the present invention constructs an indicator chain structure of the historical effective indicators to model the user browsing behavior using the Markov chain, and regards each browsing indicator as a state. The user's transition between different states is accurately captured and modeled. Compared with traditional methods, the Markov chain model can more accurately describe the dynamic characteristics of user behavior. The embodiment of the present invention calculates the state transition probability between each indicator state in the state space based on the indicator chain structure, so as to dynamically calculate the transition probability between states by analyzing a large amount of user historical browsing data. This method can adapt to changes in user behavior and ensure the accuracy and real-time nature of the transition probability, thereby improving the accuracy of the prediction. Furthermore, the embodiment of the present invention constructs a state transition matrix of the browsing behavior through the state transition probability, so as to convert the state transition probability Organized into a matrix form, an efficient state transfer matrix construction method is provided, the state transfer matrix can quickly perform matrix operations, thereby realizing efficient prediction of the user's future browsing trajectory. This efficient matrix construction method has significant advantages in processing large-scale user data. Further, the embodiment of the present invention analyzes the index browsing trajectory of the current user by using the index transfer probability, so as to not only predict the user's next behavior, but also perform multi-step prediction. By repeatedly applying the state transfer matrix, the user's future multi-step browsing trajectory can be predicted, providing a deeper behavioral insight for the recommendation system. The embodiment of the present invention recommends the current user an index of interest based on the index browsing trajectory, so as to be used for the trajectory prediction result based on the Markov chain, and the prediction result is seamlessly integrated into the recommendation system, so as to recommend the user the index data or page of the relevant industry that may be of interest in real time. Compared with the static recommendation system, the real-time recommendation system of the present invention can provide more personalized and dynamic recommendation services according to the user's current browsing state and predicted future trajectory. Therefore, the browsing trajectory analysis method, device, computer equipment and computer-readable storage medium based on the Markov chain proposed in the embodiment of the present invention can use the user's historical browsing data for accurate trajectory analysis.

[0074] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0075] In one embodiment, a browsing trajectory analysis device based on a Markov chain is provided, and the browsing trajectory analysis device based on a Markov chain corresponds to the browsing trajectory analysis method based on a Markov chain in the above embodiment. Figure 3As shown, the browsing trajectory analysis device based on Markov chain includes an index extraction module 101, a state construction module 102, a matrix construction module 103, a trajectory analysis module 104 and a result acquisition module 105. The detailed description of each functional module is as follows:

[0076] The index extraction module 101 is used to collect historical browsing records of historical users, identify index data of the historical users from the historical browsing records, record browsing states of the historical users at different time points, and extract historical valid indexes from the index data based on the browsing states;

[0077] The state construction module 102 is used to construct an indicator chain structure of the historical effective indicators, identify the indicator state of the historical user's browsing behavior from the historical effective indicators, and use the indicator state to construct the state space of the browsing behavior;

[0078] The matrix construction module 103 is used to calculate the state transition probability between each indicator state in the state space based on the indicator chain structure, and construct the state transition matrix of the browsing behavior through the state transition probability;

[0079] The trajectory analysis module 104 is used to collect the current browsing history of the current user, obtain the current valid index of the current user from the current browsing history, calculate the index transfer probability of the current valid index according to the state transfer matrix, and analyze the index browsing trajectory of the current user using the index transfer probability;

[0080] The result obtaining module 105 is used to send the interested indicators to the current user based on the indicator browsing trajectory, so as to complete the indicator browsing trajectory analysis of the current user through the interested indicators and obtain the indicator browsing trajectory analysis result of the current user.

[0081] In one embodiment of the present invention, when the indicator extraction module 101 is executed to identify the indicator data of the historical user from the historical browsing records, it is used to: divide the record indicators of the historical browsing records; obtain the timestamp of the historical browsing records; sort the historical browsing records based on the timestamp to obtain a browsing record sequence; identify target browsing records belonging to the same record indicator in the browsing record sequence; and when the timestamp of the target browsing record is a continuous timestamp, merge the target browsing records into indicator data.

[0082] In one embodiment of the present invention, when the matrix construction module 103 is executing the calculation of the state transition probability between each indicator state in the state space based on the indicator chain structure, it is used to: query the number of paths between every two historical valid indicators and the total number of departures of each historical valid indicator from the indicator chain structure; and calculate the state transition probability between each indicator state in the state space based on the path number and the total number of departures.

[0083] In one embodiment of the present invention, when executing the construction of the state transition matrix of the browsing behavior through the state transition probability, the matrix construction module 103 is used to: obtain the first transition probability belonging to the same starting point in the state transition probability; query the second transition probability belonging to the same end point in the state transition probability; use the first transition probability as the row value and the second transition probability as the column value; and construct the state transition matrix of the browsing behavior through the row value and the column value.

[0084] In one embodiment of the present invention, when the trajectory analysis module 104 calculates the indicator transfer probability of the current valid indicator according to the state transfer matrix, it is used to: construct an indicator transfer model of the current valid indicator; obtain the unknown transfer probability in the indicator transfer model; query the target transfer probability corresponding to the unknown transfer probability from the state transfer matrix; substitute the target transfer probability into the indicator transfer model to calculate the indicator transfer probability corresponding to the target transfer probability through the indicator transfer model.

[0085] In one embodiment of the present invention, when the trajectory analysis module 104 uses the indicator transfer probability to analyze the indicator browsing trajectory of the current user, it is used to: set the transfer probability threshold of the current user; when the indicator transfer probability is greater than the transfer probability threshold, obtain the indicator state sequence corresponding to the indicator transfer probability; identify the remaining indicator sequence except the current valid indicator from the indicator state sequence; and use the remaining indicator sequence as the indicator browsing trajectory.

[0086] In one embodiment of the present invention, when executing the sending of the indicator of interest to the current user based on the indicator browsing trajectory, the result obtaining module 105 is used to: obtain a first indicator page corresponding to the indicator browsing trajectory; obtain a second indicator page corresponding to the current valid indicator; calculate the page similarity between the first indicator page and the second indicator page; when the page similarity is greater than a preset similarity, use the first indicator page as the page of interest for the current user; and send the page of interest to the current user to complete the process of sending the indicator of interest to the current user.

[0087] In detail, each module in the browsing trajectory analysis device 100 based on the Markov chain in the embodiment of the present invention is used in the same manner as described above. Figure 1 The same technical means as the browsing trajectory analysis method based on Markov chain described in the text can produce the same technical effect, which will not be repeated here.

[0088] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client through a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the server side of a browsing trajectory analysis method based on a Markov chain.

[0089] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of the client side of a browsing trajectory analysis method based on a Markov chain.

[0090] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0091] Collect historical browsing records of historical users, identify index data of the historical users from the historical browsing records, record browsing status of the historical users at different time points, and extract historical valid indicators from the index data based on the browsing status;

[0092] Constructing an indicator chain structure of the historical valid indicators, identifying the indicator state of the historical user's browsing behavior from the historical valid indicators, and constructing the state space of the browsing behavior using the indicator state;

[0093] Based on the indicator chain structure, the state transition probability between each indicator state in the state space is calculated, and the state transition matrix of the browsing behavior is constructed through the state transition probability;

[0094] Collecting the current browsing history of the current user, obtaining the current valid indicator of the current user from the current browsing history, calculating the indicator transfer probability of the current valid indicator according to the state transfer matrix, and analyzing the indicator browsing trajectory of the current user by using the indicator transfer probability;

[0095] Based on the indicator browsing trajectory, the indicator of interest is sent to the current user, so as to complete the indicator browsing trajectory analysis of the current user through the indicator of interest, and obtain the indicator browsing trajectory analysis result of the current user.

[0096] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0097] Collect historical browsing records of historical users, identify index data of the historical users from the historical browsing records, record browsing status of the historical users at different time points, and extract historical valid indicators from the index data based on the browsing status;

[0098] Constructing an indicator chain structure of the historical valid indicators, identifying the indicator state of the historical user's browsing behavior from the historical valid indicators, and constructing the state space of the browsing behavior using the indicator state;

[0099] Based on the indicator chain structure, the state transition probability between each indicator state in the state space is calculated, and the state transition matrix of the browsing behavior is constructed through the state transition probability;

[0100] Collecting the current browsing history of the current user, obtaining the current valid indicator of the current user from the current browsing history, calculating the indicator transfer probability of the current valid indicator according to the state transfer matrix, and analyzing the indicator browsing trajectory of the current user by using the indicator transfer probability;

[0101] Based on the indicator browsing trajectory, the indicator of interest is sent to the current user, so as to complete the indicator browsing trajectory analysis of the current user through the indicator of interest, and obtain the indicator browsing trajectory analysis result of the current user.

[0102] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0103] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0104] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0105] The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should be included in the protection scope of the present invention. It should be noted that if software tools or components other than those of the company appear in the embodiments of this application, they are only used for illustration and do not represent actual use.

Claims

1. A browsing trajectory analysis method based on Markov chain, characterized in that: The method comprises: Collect historical browsing records of historical users, identify index data of the historical users from the historical browsing records, record browsing status of the historical users at different time points, and extract historical valid indicators from the index data based on the browsing status; Constructing an indicator chain structure of the historical valid indicators, identifying the indicator state of the historical user's browsing behavior from the historical valid indicators, and constructing the state space of the browsing behavior using the indicator state; Based on the indicator chain structure, the state transition probability between each indicator state in the state space is calculated, and the state transition matrix of the browsing behavior is constructed through the state transition probability; Collecting the current browsing history of the current user, obtaining the current valid indicator of the current user from the current browsing history, calculating the indicator transfer probability of the current valid indicator according to the state transfer matrix, and analyzing the indicator browsing trajectory of the current user by using the indicator transfer probability; Based on the indicator browsing trajectory, the indicator of interest is sent to the current user, so as to complete the indicator browsing trajectory analysis of the current user through the indicator of interest, and obtain the indicator browsing trajectory analysis result of the current user.

2. The browsing trajectory analysis method based on Markov chain according to claim 1, characterized in that: The identifying the indicator data of the historical user from the historical browsing records includes: dividing the record index of the historical browsing record; Obtaining the timestamp of the historical browsing record; Based on the timestamp, sort the historical browsing records to obtain a browsing record sequence; Identifying target browsing records belonging to the same record index in the browsing record sequence; When the timestamps of the target browsing records are continuous timestamps, the target browsing records are merged into indicator data.

3. The browsing trajectory analysis method based on Markov chain according to claim 1, characterized in that: The calculating, based on the indicator chain structure, the state transition probability between each indicator state in the state space comprises: Querying the number of paths between every two historically valid indicators and the total number of departures of each historically valid indicator from the indicator chain structure; The state transition probability between each indicator state in the state space is calculated according to the path number and the total departure number.

4. The browsing trajectory analysis method based on Markov chain according to claim 1, characterized in that: The constructing the state transition matrix of the browsing behavior by using the state transition probability includes: Obtaining a first transition probability belonging to the same starting point among the state transition probabilities; Querying the state transition probabilities for a second transition probability belonging to the same end point; Using the first transition probability as a row value and the second transition probability as a column value; A state transition matrix of the browsing behavior is constructed by using the row values ​​and the column values.

5. The browsing trajectory analysis method based on Markov chain according to claim 1, characterized in that: The calculating, according to the state transfer matrix, the indicator transfer probability of the current valid indicator comprises: Constructing an indicator transfer model of the currently effective indicator; Obtaining unknown transition probabilities in the indicator transition model; Querying the target transition probability corresponding to the unknown transition probability from the state transition matrix; The target transition probability is substituted into the indicator transition model to calculate the indicator transition probability corresponding to the target transition probability through the indicator transition model.

6. The browsing trajectory analysis method based on Markov chain according to claim 1, characterized in that: The analyzing the indicator browsing trajectory of the current user by using the indicator transition probability includes: Setting a transfer probability threshold of the current user; When the indicator transition probability is greater than the transition probability threshold, obtaining an indicator state sequence corresponding to the indicator transition probability; Identifying a remaining indicator sequence other than the currently valid indicator from the indicator state sequence; The remaining indicator sequences are used as indicator browsing tracks.

7. The browsing trajectory analysis method based on Markov chain as claimed in claim 1, characterized in that: The sending the interested indicator to the current user based on the indicator browsing track includes: Obtaining a first indicator page corresponding to the indicator browsing track; Get the second indicator page corresponding to the current valid indicator; Calculating page similarity between the first index page and the second index page; When the page similarity is greater than a preset similarity, taking the first indicator page as the page of interest to the current user; The page of interest is sent to the current user to complete the process of sending the indicator of interest to the current user.

8. A browsing trajectory analysis device based on Markov chain, characterized in that: The device comprises: An indicator extraction module is used to collect historical browsing records of historical users, identify indicator data of the historical users from the historical browsing records, record browsing states of the historical users at different time points, and extract historical valid indicators from the indicator data based on the browsing states; A state construction module, used to construct an indicator chain structure of the historical effective indicators, identify the indicator state of the historical user's browsing behavior from the historical effective indicators, and use the indicator state to construct a state space of the browsing behavior; A matrix construction module, used to calculate the state transition probability between each indicator state in the state space based on the indicator chain structure, and construct the state transition matrix of the browsing behavior through the state transition probability; A trajectory analysis module, used to collect the current browsing history of the current user, obtain the current valid index of the current user from the current browsing history, calculate the index transfer probability of the current valid index according to the state transfer matrix, and analyze the index browsing trajectory of the current user using the index transfer probability; The result obtaining module is used to send the indicators of interest to the current user based on the indicator browsing trajectory, so as to complete the indicator browsing trajectory analysis of the current user through the indicators of interest and obtain the indicator browsing trajectory analysis result of the current user.

9. A computer device, characterized in that: The computer device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the browsing trajectory analysis method based on Markov chain as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the browsing trajectory analysis method based on a Markov chain as described in any one of claims 1 to 7 is implemented.

Images