Electronic stop board health degree prediction system and prediction method

Abstract

The invention discloses an electronic stop board health degree prediction system, and the system comprises an input module which is used for obtaining variable information and calculation parameters; the calculation module is used for screening theoretical independent variables participating in regression, and regression is carried out on the prediction model of the health degree of the electronic stop board based on training data of the theoretical independent variables; the output module is used for outputting the current prediction model and a calculation result of the current prediction model about the theoretical independent variable test data; and the feedback module is connected with the output module and the input module, and after the feedback module receives negative feedback for storing the current prediction model, the input module is called to obtain the calculation parameters again. According to the invention, the health degree of the electronic stop board is predicted based on the available original independent variable information, so that workers can be helped to maintain the electronic stop board in time. The invention further provides a prediction method applied to the electronic stop board health degree prediction system.

Description

Technical Field

[0001] The present invention belongs to the technical field of electronic bus stops, and specifically relates to an electronic bus stop health prediction system and a prediction method. Background Art

[0002] At present, with the development of the urbanization process, the urban population and vehicles have increased sharply, and the road congestion situation has become increasingly severe year by year. Public transportation has the advantages of large carrying capacity, low energy consumption, energy conservation and environmental protection, and is the best choice for responding to green and environmental protection travel. Among them, the bus electronic stop is a crucial node in this bus operation loop. It solves the problem of information isolation between buses and passengers, provides a gateway for passengers to obtain information, and is a guarantee for passengers to obtain bus operation information in real time and accurately.

[0003] Therefore, how to predict the health of the electronic bus stop based on the obtained electronic bus stop information to help the staff maintain the electronic bus stop in time has become an urgent problem to be solved.

[0004] In view of this, the present invention is specifically proposed. Summary of the Invention

[0005] The first object of the present invention is to overcome the deficiencies of the prior art and provide an electronic bus stop health prediction system, which can better predict the health of the electronic bus stop through the obtained original independent variable information, so as to help the staff maintain the electronic bus stop in time.

[0006] The second object of the present invention is to provide a prediction method applied to the above-mentioned electronic bus stop health prediction system.

[0007] To achieve the first object, the basic concept of the technical solution adopted by the present invention is:

[0008] An electronic bus stop health prediction system includes:

[0009] An input module for obtaining variable information and calculation parameters. The variable information includes original independent variable information and target variable information. The calculation parameters include a selection critical value Fin for screening theoretical independent variables and a rejection critical value Fout. The theoretical independent variables participate in the regression prediction model;

[0010] A calculation module for screening the theoretical independent variables participating in the regression and regressing a prediction model for the health of the electronic bus stop based on the training data of the theoretical independent variables;

[0011] An output module for outputting the current prediction model and its calculation results for the test data of the theoretical independent variables;

[0012] A feedback module, connected to the output module and the input module. After receiving a negative feedback for saving the current prediction model, the feedback module calls the input module to re-obtain calculation parameters.

[0013] Further, the original independent variable information includes one or more of ARM processor information, central controller information, display screen information, and fan information.

[0014] Further, the ARM processor information and fan information obtained by the input module are in the form of binary variables, and the central controller information and display screen information are in the form of continuous variables;

[0015] The target variable information obtained by the input module is in the form of binary variables.

[0016] To achieve the second objective, the basic concept adopted by the present invention is:

[0017] A prediction method applied to the above-mentioned electronic signboard health prediction system includes the following steps:

[0018] S1. Obtain the original independent variable information and the target variable information;

[0019] S2. Divide the original independent variable information into training data and test data, and obtain the selected critical value Fin and the rejection critical value Fout to screen the theoretical independent variables for the regression prediction model;

[0020] S3. Based on the training data of the theoretical independent variables, regress the prediction model of the electronic signboard health, and output the current model and its calculation results regarding the test data of the theoretical independent variables;

[0021] S4. Ask the user whether to save the current model. If not, jump to step S2 to re-obtain one or more parameters of the ratio relationship between the training data and the test data, the selected critical value Fin, and the rejection critical value Fout.

[0022] Further, step S1 further includes performing a standardization process on the continuous variables, and its calculation formula is:

[0023] x = (x - min(x)) / (max(x) - min(x)).

[0024] Further, when participating in the calculation, 0 of the target variable is changed to 0.1, 1 is changed to 0.9, and it is transformed into a target variable calculated value that conforms to the normal distribution. The calculation formula of the target variable calculated value is log(p / (1 - p));

[0025] Where p is the value of the target variable after transformation.

[0026] Further, the theoretical independent variables screened in step S2 include:

[0027] S21. Calculate the first F-test value of all original independent variables and determine whether it is greater than the selected critical value Fin. If so, the corresponding original independent variable is selected;

[0028] If not, the corresponding original independent variable is excluded;

[0029] S22. Calculate the second F-test value of all selected original independent variables and determine whether it is greater than the exclusion critical value Fout. If so, define the corresponding original independent variable as a theoretical independent variable;

[0030] If not, the corresponding original independent variable is excluded.

[0031] Further, the calculation formula of the first F-test value is:

[0032]

[0033] The calculation formulas of the matrices Syy, Sxy, and Sxx are:

[0034]

[0035] where j = 1, 2... m;

[0036]

[0037] where n is the number of original independent variables, m is the number of rows of the original independent variable training data, x is the theoretical independent variable participating in the calculation, and y is the calculated value of the target variable.

[0038] Further, the calculation formula of the second F-test value is:

[0039]

[0040] where r is the number of original independent variables selected in step S22.

[0041] Further, the calculation formula of the electronic stop board health degree z is:

[0042]

[0043] where β is the regression coefficient corresponding to the theoretical independent variable, and x is the corresponding theoretical independent variable;

[0044] The output value of the target variable is in the form of a binary variable. When the electronic stop board health degree z is less than or equal to 0.5, the output is 0;

[0045] When the health degree z of the electronic stop sign is greater than 0.5, the output is 1.

[0046] After adopting the above technical solution, the present invention has the following beneficial effects compared with the prior art:

[0047] The present invention predicts the health degree of the electronic stop sign based on the available original independent variable information, which can help the staff to maintain the electronic stop sign in a timely manner; on the other hand, the staff can also adjust the relevant parameters participating in the regression model based on the prediction result of the current model output, so as to obtain the prediction model with the best prediction result.

[0048] The following further describes in detail the specific embodiments of the present invention with reference to the accompanying drawings. Description of the Drawings

[0049] As a part of the present invention, the accompanying drawings are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention, but do not constitute an improper limitation to the present invention. Obviously, the accompanying drawings in the following description are only some embodiments, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0050] In the accompanying drawings:

[0051] Figure 1 is a schematic flow chart of the method for predicting the health degree of the electronic stop sign of the present invention;

[0052] Figure 2 is a schematic flow chart of screening the theoretical independent variables of the present invention.

[0053] It should be noted that these drawings and text descriptions are not intended to limit the scope of the concept of the present invention in any way, but to illustrate the concept of the present invention to those skilled in the art by referring to specific embodiments. Specific Embodiments

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0055] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", "front", "rear", "left", "right", "vertical", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0056] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0057] The present invention provides an electronic signboard health prediction system, which predicts the health of the electronic signboard based on the available original independent variable information, so as to help the staff maintain the electronic signboard in a timely manner.

[0058] Specifically, the electronic signboard health prediction system of the present invention includes:

[0059] An input module for obtaining variable information and calculation parameters, where the variable information includes original independent variable information and target variable information, and the calculation parameters include a selection critical value Fin and a rejection critical value Fout for screening theoretical independent variables. The theoretical independent variables are the independent variables selected from the original independent variables for the regression prediction model.

[0060] A calculation module for screening the theoretical independent variables participating in the regression and regressing the prediction model of the electronic signboard health based on the training data of the theoretical independent variables.

[0061] An output module for outputting the current prediction model and its calculation results regarding the test data of the theoretical independent variables.

[0062] A feedback module is connected to the output module and the input module. After receiving a negative feedback for saving the current prediction model, the feedback module calls the input module to re-obtain the calculation parameters. That is, when the user is not satisfied with the calculation results of the current prediction model, the user can choose not to save the current prediction model. At this time, the feedback module receives the negative feedback for saving the current prediction model and calls the input module to re-regress the prediction model in the form of re-obtaining the calculation parameters to obtain better prediction results.

[0063] Furthermore, the original independent variable information includes one or more of ARM processor information, central controller information, display screen information, and fan information.

[0064] For the convenience of calculation, in the present invention, the ARM processor information and fan information obtained by the input module are in the form of binary variables, and the central controller information and display screen information are in the form of continuous variables.

[0065] On the other hand, the target variable information obtained by the input module is also in the form of binary variables.

[0066] Specifically, the ARM processor information obtained in the present invention is in the form of binary variables, including two values, 0 and 1, where 0 represents zero repairs and 1 represents one or more repairs.

[0067] Regarding the fan information, considering that it may include the situations of zero repairs, one repair, two repairs, and three or more repairs, the present invention correspondingly sets its variable names as x1, x2, x3, and x4. At this time, 0 of the binary variable represents no, and 1 represents yes.

[0068] Regarding the target variable information, it includes two values, 0 and 1, where 1 of the target variable preferably represents a healthy product, and correspondingly, 0 preferably represents a product that needs maintenance.

[0069] In an embodiment of the present invention, the original independent variable information and the target variable information are as shown in the following table:

[0070]

[0071] As Figure 1-2 shown, the present invention also provides a prediction method applied to the above electronic signboard health prediction system, including the following steps:

[0072] S1. Obtain the original independent variable information and the target variable information;

[0073] S2. Divide the original independent variable information into training data and test data, and obtain the selected critical value Fin and the rejection critical value Fout to screen the theoretical independent variables for the regression prediction model;

[0074] S3. Based on the training data of the theoretical independent variables, regress the prediction model of the electronic signboard health, and output the current model and its calculation results for the test data of the theoretical independent variables;

[0075] S4. Ask the user whether to save the current model. If not, jump to step S2 to re-obtain one or more parameters among the proportional relationship between the training data and the test data, the selected critical value Fin, and the rejection critical value Fout.

[0076] In the present invention, the staff can regress the prediction model of the electronic signboard health based on all the available electronic signboard information, so as to timely maintain the electronic signboard according to the calculation results of the prediction model. When the staff is not satisfied with the prediction results of the current prediction model, they can also timely adjust the regressed prediction model by adjusting parameters such as the proportional relationship between the original independent variable training data and the test data, the selected critical value Fin, and the rejection critical value Fout, and then obtain the optimal prediction model.

[0077] In one embodiment of the present invention, considering that the centralized controller information and the display screen information are in the form of continuous variables, for the convenience of calculation, this embodiment performs a 0-1 normalization process on them. Specifically, the calculation formula is:

[0078] x=(x-min(x)) / (max(x)-min(x)).

[0079] On the other hand, considering that linear regression is the predicted value of a new data, it is actually the expectation of normal distribution. Therefore, in order to ensure the regression effect, the calculated value of the target variable for regression must also conform to the normal distribution.

[0080] Specifically, in this embodiment, the 0 of the target variable involved in the calculation is first converted to 0.1, and 1 is converted to 0.9, and then the converted target variable value p is transformed into a target variable calculation value that conforms to the normal distribution, and its calculation formula is log(p / (1-p)).

[0081] Through the above operations, the original independent variable information and target variable calculation value obtained can be applied to the regression prediction model.

[0082] In this embodiment, in order to test whether the prediction model obtained by regression is accurate, the original independent variable information is divided into training data and test data, wherein the training data is used for the regression prediction model, and the prediction model can output the calculation results according to the original independent variable information of the test data, and the accuracy of the current prediction model can be judged by comparing the calculation results with the actual results of the test data, and then the user can judge whether to adopt the current prediction model or re-regress according to the comparison results. It should be understood that the staff can define any original independent variable information as training data or test data.

[0083] In another embodiment of the present invention, step S1 also includes cleaning the original independent variable information and target variable information. For example, in the case of missing data or abnormal data, this embodiment can choose to eliminate relevant data or replace the missing or abnormal part of the data by taking the average value.

[0084] like Figure 2 As shown, in another embodiment of the present invention, the theoretical independent variables screened for the regression prediction model in step S2 include:

[0085] S21, calculate the first F test value of all original independent variables and determine whether it is greater than the selection critical value Fin, if so, the corresponding original independent variable is selected;

[0086] If not, the corresponding original independent variable is eliminated;

[0087] S22. Calculate the second F-test value of all selected original independent variables and determine whether it is greater than the elimination critical value Fout. If so, define the corresponding original independent variable as the theoretical independent variable;

[0088] If not, the corresponding original independent variable is eliminated.

[0089] Among them, the selection critical value Fin needs to be greater than or equal to 1.5 and less than or equal to 4; and the elimination critical value Fout needs to be greater than or equal to 3 and less than or equal to 10.

[0090] Furthermore, the selection critical value Fin participating in the calculation needs to be less than or equal to the elimination critical value Fout. For example, the selection critical value Fin can be set to 2.5, and the elimination critical value can be set to 3.

[0091] When the staff needs to adjust the prediction model, they can adjust the theoretical independent variables participating in the regression model by adjusting the selection critical value Fin and the elimination critical value Fout, and then adjust the output prediction model.

[0092] The calculation formula for the first F-test value is:

[0093]

[0094] Among them, the calculation formulas for the matrices Syy, Sxy, and Sxx are:

[0095]

[0096] Among them, j = 1, 2... m;

[0097] Among them, i, j = 1, 2... m;

[0098] At this time, the matrix

[0099] Among them, n is the number of original independent variables, m is the number of rows of training data in the original independent variables, x is the theoretical independent variable participating in the calculation, and y is the calculated value of the target variable participating in the calculation.

[0100] Assume that when the t variable is selected, the elimination transformation of the matrix S can be obtained as:

[0101]

[0102]

[0103] At this time, the last row of data in matrix S corresponds to the regression coefficients of the theoretical independent variables. However, since the selected original independent variables have not been further screened by the elimination critical value Fout, the amount of data in the last row of the current matrix S is greater than or equal to the number of regression coefficients of the final theoretical independent variables.

[0104] The calculation formula for the second F-test value is:

[0105]

[0106] where r is the number of original independent variables selected in step S22.

[0107] Through the above process, the theoretical independent variables participating in the regression model can be determined. In the present invention, the number of theoretical independent variables in the prediction model is less than or equal to the number of original independent variables obtained. For example, although the original independent variables include 7 items such as ARM processor information, central controller information, display screen information, x1, x2, x3, and x4, after screening by the selection critical value Fin and the elimination critical value Fout, only two original independent variables, namely ARM processor information and central controller information, may remain. These two original independent variables are the theoretical independent variables participating in the regression. In other words, under the screening of the current selection critical value Fin and elimination critical value Fout, the influence of original independent variables such as display screen information and fan information on the health of the electronic signboard can be ignored.

[0108] After determining the theoretical independent variables, their corresponding regression coefficients can be determined. Specifically, in this embodiment, the data in the last row of matrix S after performing the elimination transformation is the regression coefficient of the theoretical independent variables. That is, after screening out the theoretical independent variables through the elimination critical value Fout, the regression coefficients corresponding one-to-one to the theoretical independent variables can be further screened out from the data in the last row of matrix S after performing the elimination transformation.

[0109] At this time, the calculation formula for the health z of the electronic signboard is:

[0110]

[0111] where β is the regression coefficient corresponding to the theoretical independent variable, and x is the corresponding theoretical independent variable.

[0112] Furthermore, in this embodiment, the output value of the target variable is also in the form of a binary variable. When the calculated value of the health z of the electronic signboard is less than or equal to 0.5, the output is 0; when it is greater than 0.5, the output is 1.

[0113] For example, when the calculation result of the above calculation formula is 0.6 or 0.8, the output is 1. At this time, 1 indicates that the current electronic signboard is a healthy product; when the calculation result is 0.4, the output is 0. At this time, 0 indicates that the current electronic signboard is a product in need of maintenance, and the staff needs to pay attention.

[0114] The accuracy of the current prediction model can be obtained by comparing the actual value and the calculated value of the test data. The staff can determine whether to adopt the current prediction model according to the output accuracy. If the current model is not adopted, the relevant parameters can be retrieved again in step S2. In addition to retrieving the inclusion threshold Fin and the exclusion threshold Fout again, under the adjustment of the staff, in this embodiment, the ratio of the training data to the test data in the original independent variable information can also be retrieved again. For example, when the original independent variable information is 4 items, the ratio of the training data to the test data can be 1:3, 2:2, or 3:1.

[0115] In another embodiment of the present invention, after obtaining the original independent variable information, based on the inclusion threshold Fin and the exclusion threshold Fout set by the staff, this embodiment can continuously compare the calculated value of the test data calculated by the current prediction model with its actual value, and directly output the prediction model with the highest accuracy.

[0116] Or after the staff sets the range of the inclusion threshold Fin and the exclusion threshold Fout, this embodiment can directly output the prediction model of the electronic sign health degree with the highest accuracy within the current inclusion threshold Fin and the exclusion threshold Fout.

[0117] The above are only the preferred embodiments of the present invention, and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art of this patent, without departing from the scope of the technical solution of the present invention, can make some changes or modifications to the above-mentioned technical content prompted to be equivalent embodiments of equivalent changes. The implementation schemes in the above embodiments can also be further combined or replaced. However, as long as it does not depart from the content of the technical solution of the present invention, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention still fall within the scope of the present invention.

Claims

1. An electronic signboard health prediction system, characterized in that, it includes: An input module for obtaining variable information and calculation parameters. The variable information includes original independent variable information and target variable information. The calculation parameters include a selection critical value Fin and a rejection critical value Fout for screening theoretical independent variables. The theoretical independent variables participate in the regression prediction model; A calculation module for screening the theoretical independent variables participating in the regression and regressing the prediction model of the electronic signboard health based on the training data of the theoretical independent variables; An output module for outputting the current prediction model and its calculation results regarding the test data of the theoretical independent variables; A feedback module connected to the output module and the input module. After receiving a negative feedback to save the current prediction model, the feedback module calls the input module to re-obtain the calculation parameters.

2. An electronic signboard health prediction system according to claim 1, characterized in that, The original independent variable information includes one or more of ARM processor information, centralized controller information, display screen information, and fan information.

3. An electronic signboard health prediction system according to claim 2, characterized in that, The ARM processor information and fan information obtained by the input module are in the form of binary variables, and the centralized controller information and display screen information are in the form of continuous variables; The target variable information obtained by the input module is in the form of binary variables.

4. A prediction method applied to the electronic signboard health prediction system according to any one of claims 1-3, characterized in that, it includes the following steps: S1. Obtain the original independent variable information and target variable information; S2. Divide the original independent variable information into training data and test data, and obtain the selection critical value Fin and the rejection critical value Fout to screen the theoretical independent variables for the regression prediction model; S3. Regress the prediction model of the electronic signboard health based on the training data of the theoretical independent variables, and output the current model and its calculation results regarding the test data of the theoretical independent variables; S4. Ask the user whether to save the current model. If not, jump to step S2 to re-obtain one or more parameters among the ratio relationship between the training data and the test data, the selection critical value Fin, and the rejection critical value Fout.

5. A prediction method according to claim 4, characterized in that, Step S1 further includes performing a standardization process on the continuous variable, and its calculation formula is: x = (x - min(x)) / (max(x) - min(x)).

6. A prediction method according to claim 4, characterized in that, When participating in the calculation, 0 of the target variable is changed to 0.1, 1 is changed to 0.9, and it is transformed into a target variable calculated value conforming to the normal distribution. The calculation formula of the target variable calculated value is log(p / (1 - p)); where p is the value of the target variable after transformation.

7. A prediction method according to claim 4, characterized in that, The screening of the theoretical independent variables for the regression prediction model in step S2 includes: S21. Calculate the first F - test value of all original independent variables and determine whether it is greater than the selected critical value Fin. If so, the corresponding original independent variable is selected; If not, the corresponding original independent variable is excluded; S22. Calculate the second F - test value of all selected original independent variables and determine whether it is greater than the exclusion critical value Fout. If so, define the corresponding original independent variable as the theoretical independent variable; If not, the corresponding original independent variable is excluded.

8. A prediction method according to claim 7, wherein, The calculation formula for the first F - test value is: The calculation formulas for the matrices Syy, Sxy, and Sxx are: where j = 1, 2... m; where i, j = 1, 2... m; where n is the number of original independent variables, m is the number of rows of the training data of the original independent variables, x is the theoretical independent variable participating in the calculation, and y is the calculated value of the target variable.

9. A prediction method according to claim 8, wherein, The calculation formula for the second F - test value is: where r is the number of original independent variables selected in step S22.

10. A prediction method according to any one of claims 4 - 9, wherein, The calculation formula for the health degree z of the electronic stop board is: where β is the regression coefficient corresponding to the theoretical independent variable, and x is the corresponding theoretical independent variable; The output value of the target variable is in the form of a binary variable. When the health degree z of the electronic stop board is less than or equal to 0.5, the output is 0; When the health degree z of the electronic stop board is greater than 0.5, the output is 1.

Images