Wine market prediction method and device

Abstract

The invention relates to the technical field of computers, and particularly provides a wine market prediction method and device, and the method is based on a weighted Markov chain, and comprises the following steps: S1, formulating a sales grading standard of a wine market; s2, detecting the Markov property of the wine market sales data; s3, calculating the weighted transition probability of the Markov chain; s4, analyzing wine market state prediction; s5, constructing a wine market prediction model; and S6, iterating the wine market prediction model. Compared with the prior art, the method can more accurately predict the possible sales state of the wine market in the future.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular provides a wine market forecasting method and device. Background Art

[0002] With the development of information technology, the prediction of market changes has attracted more and more attention from various enterprises. The current prediction of changes in the wine market is mainly based on the prediction of simple mathematical models, which use the market sales data, fit the sales curve, and predict the sales trend of the market. The data processing is too simple, the degree of information mining is low, and the prediction result is low in accuracy, which cannot accurately predict changes in the wine market. Summary of the invention

[0003] The present invention aims at solving the above-mentioned deficiencies of the prior art and provides a wine market forecasting method with strong practicability.

[0004] A further technical task of the present invention is to provide a wine market forecasting device that is rationally designed, safe and applicable.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] A wine market prediction method based on a weighted Markov chain has the following steps:

[0007] S1. Formulate the classification standards for alcoholic beverage sales;

[0008] S2, test the Markov property of wine market sales data;

[0009] S3, calculating the weighted transition probability of the Markov chain;

[0010] S4. Analyze and forecast the status of the wine market;

[0011] S5. Construct a wine market forecasting model;

[0012] S6. Iterate the wine market prediction model.

[0013] Furthermore, in step S1, the wine sales data is selected, the total sales of each month is counted, the sample average is used as the classification center, the sample mean square error is used as the variation scale, and a set of classification standard coefficients is set. When 2n levels are divided and n-1 standard coefficients are set, the classification standard formula can be expressed as:

[0014]

[0015] Among them, the Grade function is a piecewise function, which takes different state values ​​when the sales volume is in different segments. is the average value of the sample, s is the sample mean, 2n is the number of sample classifications, that is, the number of states of the Markov chain, k1k2…k n-1 is the classification standard coefficient.

[0016] Further, in step S2, the monthly market sales data of alcoholic beverages is represented as {X n}, if the sequence can satisfy the Markov test, then a Markov chain is constructed to predict {X n}Change pattern;

[0017] Use f ij represents the sequence X1,X2,…X n The frequency of starting from state i and transferring to state j in one step is calculated by taking all f ij The matrix with elements is the transition frequency matrix. The edge frequency is obtained by dividing the sum of each column of the transition frequency matrix by the sum of each row and column, and it is used as the estimate of the edge frequency, denoted as Right now Assume that there are m types of states in the wine market sales sequence;

[0018] According to the knowledge of probability theory, the statistic The degree of freedom is (m-1) 2 χ 2 Distribution, among which, Represents the transition probability matrix of the wine market sales series. Given the significance level α, the distribution points are obtained The value of the statistic χ is obtained by calculating 2 If the value of Then the sequence {X n}According to Markov property, the wine market sales series is treated as a Markov chain;

[0019] The sequence is processed according to the wine market sales grading standard, the grade corresponding to the monthly sales of the wine market is calculated, the state space of the Markov chain is constructed, and the state of each indicator value in the research sequence is determined.

[0020] Furthermore, in step S3, it includes:

[0021] S3.1. Calculate the autocorrelation coefficient of step S1 to the mth order:

[0022]

[0023] Among them, r k represents the k-order autocorrelation coefficient, X i represents the index value of the i-th period, represents the average value of the sequence, n represents the length of the sequence, and m represents the maximum order that needs to be calculated during prediction, usually the number of states of the sequence;

[0024] S3.2, calculate the weighting coefficients of step S1 to the mth order:

[0025]

[0026] Among them, W k represents the weighting coefficient of order k;

[0027] S3.3, calculate the transition probability matrix from step S1 to the mth order:

[0028] p ij (k) =P{X(m+k)=i m+k |X(m)=i m},i,j∈E; (4)

[0029] Among them, p ij (k) It indicates that the system is in state i, and after k steps of state transition, the probability of reaching state j, m represents the current time, k represents the transition time, i m represents the state of the system at time m, i m+k Represents the state of the system after k steps of transfer.

[0030] Furthermore, in step S4, the sales volume of the previous several months is used as the initial state, and the probability of state i in the month can be predicted by using the corresponding transition probability according to the step length k required to transfer to the predicted month: p i (k) ;

[0031] The predicted probabilities of each state are weighted using the corresponding weights, and finally the predicted probability of the system being in state i is obtained:

[0032]

[0033] Among them, p i Represents the probability of the system transferring from a certain state to state i;

[0034] At this time, max{p i The state value corresponding to} is the most likely state of the wine market in the next month.

[0035] Furthermore, in step S5, the sample mean and sample mean square error are first calculated:

[0036]

[0037] Divide the samples into 6 standard grades, set the wine market classification standard parameter k, and divide the samples according to formula (1); divide the samples according to the classification standard to generate sales status; calculate the autocorrelation coefficients of order 1 to 6 according to formula (2), and then use formula (3) to calculate the weights corresponding to the autocorrelation coefficients of order 6 respectively;

[0038] According to formula (4), the one-step transition probability matrix of the Markov chain of each step length is obtained, and the market of future months is predicted based on the wine market sales data of previous months.

[0039] Furthermore, in step S6, the data of future months are added to the calculation process in step S5 for predicting the wine market sales in future months. Similarly, the actual sales data generated in future months are iteratively added to the wine market prediction model to test and iteratively optimize the model.

[0040] A wine market prediction device, comprising: at least one memory and at least one processor;

[0041] The at least one memory is used to store a machine-readable program;

[0042] The at least one processor is used to call the machine-readable program to execute a wine market forecasting method.

[0043] Compared with the prior art, the wine market forecasting method and device of the present invention have the following outstanding beneficial effects:

[0044] The present invention maps the sales data to the corresponding sales status by formulating the sales classification standard of the wine market; then performs a Markov property test on the sales data of the wine market to ensure the feasibility and reliability of the prediction model; uses the idea of ​​weighted Markov chain to calculate the sales data of the wine market to obtain the transition probability matrix of the wine market; derives the wine market status prediction formula through predictive analysis of the wine market status; uses actual data to construct a prediction model for the wine market, and tests and iteratively optimizes the model, so as to more accurately predict the possible sales status of the wine market in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Attached Figure 1 It is a flow chart of a wine market forecasting method. DETAILED DESCRIPTION

[0047] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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.

[0048] A best embodiment is given below:

[0049] like Figure 1 As shown, a wine market prediction method in this embodiment is based on a weighted Markov chain and has the following steps:

[0050] S1. Formulate the classification standards for alcoholic beverage sales;

[0051] Select the wine sales data, count the total sales of each month, take the sample average as the classification center, the sample mean square error as the change scale, and set the classification standard coefficient set. When divided into 2n levels and set n-1 standard coefficients, the classification standard formula can be expressed as:

[0052]

[0053] Among them, the Grade function is a piecewise function, which takes different state values ​​when the sales volume is in different segments. is the average value of the sample, s is the sample mean, 2n is the number of sample classifications, that is, the number of states of the Markov chain, k1k2…k n-1 is the classification standard coefficient.

[0054] S2, test the Markov property of wine market sales data;

[0055] Monthly market sales data of alcoholic beverages are expressed as {X n}, if the sequence can satisfy the Markov test, then a Markov chain can be constructed to predict {X n}'s changing rules.

[0056] Among them, the Markov property test is used to determine whether a sequence has the Markov property, that is, whether the future state of the sequence depends only on the current state and has nothing to do with the previous state. This test is usually used to verify whether a random process can be reasonably modeled as a Markov chain.

[0057] Use f ij represents the sequence X1,X2,…X n The frequency of starting from state i and transferring to state j in one step is calculated by taking all f ijThe matrix with elements is the transition frequency matrix. The edge frequency is obtained by dividing the sum of each column of the transition frequency matrix by the sum of each row and column, and it is used as the estimate of the edge frequency, denoted as Right now (Assume that there are m types of states in the wine market sales sequence).

[0058] According to the knowledge of probability theory, statistics The degree of freedom is (m-1) 2 χ 2 Distribution, where The transition probability matrix represents the sales series of the wine market. Given a significance level α, the distribution point can be obtained by looking at the chi-square distribution table. The value of the statistic χ is obtained by calculating 2 If the value of Then the sequence {X n}According to the Markov property, the wine market sales series can be treated as a Markov chain.

[0059] The sequence is processed according to the wine market sales grading standard, the grade corresponding to the monthly sales of the wine market is calculated, the state space of the Markov chain is constructed, and the state of each indicator value in the research sequence is determined.

[0060] S3, calculating the weighted transition probability of the Markov chain;

[0061] include:

[0062] S3.1. Calculate the autocorrelation coefficient of step S1 to the mth order:

[0063]

[0064] Among them, r k represents the k-order autocorrelation coefficient, X i represents the index value of the i-th period, represents the average value of the sequence, n represents the length of the sequence, and m represents the maximum order that needs to be calculated during prediction, usually the number of states of the sequence;

[0065] S3.2, calculate the weighting coefficients of step S1 to the mth order:

[0066]

[0067] Among them, W k represents the weighting coefficient of order k;

[0068] S3.3, calculate the transition probability matrix from step S1 to the mth order:

[0069] p ij (k) =P{X(m+k)=im+k |X(m)=i m}, i,j∈E; (4)

[0070] Among them, p ij (k) It indicates that the system is in state i, and after k steps of state transition, the probability of reaching state j, m represents the current time, k represents the transition time, i m represents the state of the system at time m, i m+k Represents the state of the system after k steps of transfer.

[0071] S4. Analyze and forecast the status of the wine market;

[0072] Taking the sales volume of the previous several months as the initial state, according to the step length k required to transfer to the predicted month, using the corresponding transition probability, the probability of state i in that month can be predicted: p i (k) ;

[0073] The predicted probabilities of each state are weighted using the corresponding weights, and finally the predicted probability of the system being in state i is obtained:

[0074]

[0075] Among them, p i Represents the probability of the system transferring from a certain state to state i;

[0076] At this time, max{p i The state value corresponding to} is the most likely state of the wine market in the next month.

[0077] S5. Construct a wine market forecasting model;

[0078] The prediction model uses the recent sales of a certain brand of wine as an example for prediction analysis. First, the sample mean and sample mean square error are calculated: Ten thousand yuan / month, Ten thousand yuan / month.

[0079] The samples are divided into 6 standard grades (n = 3), and the wine market classification standard parameter k is set to [1.0, 0.5]. According to formula (1), the sample classification results are shown in the table:

[0080] Table 1 Wine market classification standards

[0081]

[0082] The samples are divided according to the grading standards to generate sales status. The mapping results are shown in the table:

[0083] Table 2 Sales volume sequence and status of the alcoholic beverage market in the past 20 months

[0084]

[0085] According to formula (2), the autocorrelation coefficients of order 1 to 6 are calculated, and then the weights corresponding to the autocorrelation coefficients of order 6 are calculated using formula (3). The results are shown in the table:

[0086] Table 3 Autocorrelation coefficients and their weights

[0087]

[0088] According to formula (4), the one-step transition probability matrix of the Markov chain of each step length can be obtained as follows:

[0089]

[0090] Based on the sales data of a certain brand of alcoholic beverages in the 15th to 20th months, the market forecast for the 21st month is as follows:

[0091] Table 4: Forecast calculation table of wine market in the 21st month

[0092]

[0093] According to the above calculation results, the probability of conversion to state 5 is the highest in the 21st month. Therefore, it can be predicted that the sales volume of the wine market in the 21st month should be between 207,200 yuan and 305,500 yuan.

[0094] S6, iterative wine market prediction model;

[0095] The actual data shows that the alcohol sales in the 21st month is RMB 253,200, which is consistent with the prediction results of the above model. The data of the 21st month is added to the calculation process in step 5 to predict the alcohol market sales in subsequent months. Similarly, the actual sales data generated in subsequent months is iteratively added to the alcohol market prediction model to test and iteratively optimize the model.

[0096] This application uses Markov chain to predict changes in the wine market. Compared with the predictions of other probability statistics algorithms, Markov chain can more effectively capture the changing trends in the data sequence, especially when processing data with Markov properties, where the future state is only related to the current state and not to the past state. This feature avoids a series of difficulties encountered by other prediction methods when collecting historical data, making it occupy an important position both in theory and application.

[0097] Based on the above method, a wine market prediction device in this embodiment includes: at least one memory and at least one processor;

[0098] The at least one memory is used to store a machine-readable program;

[0099] The at least one processor is used to call the machine-readable program to execute a wine market forecasting method.

[0100] The above-mentioned specific implementations are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementations. Any technical solutions that conform to the above-mentioned specific implementations of the present invention and any appropriate changes or substitutions made by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.

[0101] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A wine market forecasting method, characterized in that: Based on the weighted Markov chain, the steps are as follows: S1. Formulate the classification standards for alcoholic beverage sales; S2, test the Markov property of wine market sales data; S3, calculating the weighted transition probability of the Markov chain; S4. Analyze and forecast the status of the wine market; S5. Construct a wine market forecasting model; S6. Iterate the wine market prediction model.

2. A wine market forecasting method according to claim 1, characterized in that: In step S1, the wine sales data is selected, the total sales of each month is counted, the sample average is used as the classification center, the sample mean square error is used as the variation scale, and a set of classification standard coefficients is set. When 2n levels are divided and n-1 standard coefficients are set, the classification standard formula can be expressed as: Among them, the Grade function is a piecewise function, which takes different state values ​​when the sales volume is in different segments. is the average value of the sample, s is the sample mean, 2n is the number of sample classifications, that is, the number of states of the Markov chain, k1k2…k n-1 is the classification standard coefficient.

3. A wine market forecasting method according to claim 2, characterized in that: In step S2, the monthly market sales data of alcoholic beverages is represented as {X n }, if the sequence can satisfy the Markov test, then a Markov chain is constructed to predict {X n }The changing rules of Use f ij represents the sequence X1,X2,…X n The frequency of starting from state i and transferring to state j in one step is calculated by taking all f ij The matrix with elements is the transition frequency matrix. The edge frequency is obtained by dividing the sum of each column of the transition frequency matrix by the sum of each row and column, and it is used as the estimate of the edge frequency, denoted as Right now Assume that there are m types of states in the wine market sales sequence; According to the knowledge of probability theory, the statistic The degree of freedom is (m-1) 2 χ 2 Distribution, among which, Represents the transition probability matrix of the wine market sales series. Given the significance level α, the distribution points are obtained The value of χ is obtained by calculating the statistic 2 If the value of Then the sequence {X n }According to Markov property, the wine market sales series is treated as a Markov chain; The sequence is processed according to the wine market sales grading standard, the grade corresponding to the monthly sales of the wine market is calculated, the state space of the Markov chain is constructed, and the state of each indicator value in the research sequence is determined.

4. A wine market forecasting method according to claim 3, characterized in that: In step S3, it includes: S3.

1. Calculate the autocorrelation coefficient of step S1 to the mth order: Among them, r k represents the k-order autocorrelation coefficient, X i represents the index value of the i-th period, represents the average value of the sequence, n represents the length of the sequence, and m represents the maximum order that needs to be calculated during prediction, usually the number of states of the sequence; S3.2, calculate the weighting coefficients of step S1 to the mth order: Among them, W k represents the weighting coefficient of order k; S3.3, calculate the transition probability matrix from step S1 to the mth order: p ij (k) =P{X(m+k)=i m+k |X(m)=i m },i,j∈E; (4) Among them, p ij (k) It indicates that the system is in state i, and after k steps of state transition, the probability of reaching state j, m represents the current time, k represents the transition time, i m represents the state of the system at time m, i m+k Represents the state of the system after k steps of transfer.

5. A wine market forecasting method according to claim 4, characterized in that: In step S4, the sales volume of the previous months is used as the initial state, and the probability of state i in the month can be predicted by using the corresponding transition probability according to the step length k required to transfer to the predicted month: p i (k) ; The predicted probabilities of each state are weighted using the corresponding weights, and finally the predicted probability of the system being in state i is obtained: Among them, p i Represents the probability of the system transferring from a certain state to state i; At this time, max{p i The state value corresponding to} is the most likely state of the wine market in the next month.

6. A wine market forecasting method according to claim 4, characterized in that: In step S5, the sample mean and sample mean square error are first calculated: Divide the samples into 6 standard grades, set the wine market classification standard parameter k, and divide the samples according to formula (1); divide the samples according to the classification standard to generate sales status; calculate the autocorrelation coefficients of order 1 to 6 according to formula (2), and then use formula (3) to calculate the weights corresponding to the autocorrelation coefficients of order 6 respectively; According to formula (4), the one-step transition probability matrix of the Markov chain of each step length is obtained, and the market of future months is predicted based on the wine market sales data of previous months.

7. A wine market forecasting method according to claim 6, characterized in that: In step S6, the data of future months are added to the calculation process in step S5 for predicting the wine market sales in future months. Similarly, the actual sales data generated in future months are iteratively added to the wine market prediction model to test and iteratively optimize the model.

8. A wine market forecasting device, characterized in that: include: at least one memory and at least one processor; The at least one memory is used to store a machine-readable program; The at least one processor is configured to call the machine-readable program to execute the method according to any one of claims 1 to 7.

Images