Parking lot income prediction method

Abstract

The invention discloses a parking lot income prediction method, and relates to the technical field of digital parking operation, and the method comprises the steps: carrying out the detailed analysis of the historical parking data of a parking lot, and obtaining different operation day types, such as usual, weekend, holidays and festivals, and restaurant activities; taking the average charging traffic flow and the average charging duration of different operation day types as income parameters for future income measurement and calculation; a prediction model for predicting parking lot income in a future set time period can be constructed based on a deep learning algorithm; according to the prediction method, accurate and specific time prediction of a future set time period can be realized for different parking lots, and a parking lot manager can better plan parking resources. The prediction result can help the financial department to make a budget more accurately, carry out cost control and financial planning, and reduce the operation risk. Through income prediction, market trends and consumer behaviors can be analyzed, and marketing strategies and price strategies can be adjusted in time to attract more users.

Description

Technical Field

[0001] The present invention relates to the technical field of digital parking operation, and in particular to a parking lot revenue prediction method. Background Art

[0002] The prediction of parking lot revenue is of great significance to the business decision-making of parking lots. At present, the prediction technology of parking lot revenue is mainly based on the linear statistical analysis of historical data;

[0003] Take the charging rules of a certain parking lot as an example: temporary parking fee standard: 4 yuan / hour; monthly rental fee standard: 100 yuan / month for internal employees; 150 yuan / month for merchant vehicles; 2 hours free for purchases over 50 yuan and less than 100 yuan; 3 hours free for purchases over 100 yuan; 4 hours free with movie tickets.

[0004] The traditional method of predicting parking revenue is: number of parking lots * 365 days * 4 yuan / day * full parking rate. But how many cars are actually in the free period? The proportion of free cars varies greatly due to the different customer numbers and consumption types on weekends, holidays, and weekdays. The traditional linear calculation formula cannot accurately predict.

[0005] Therefore, it is urgent to design a method to intelligently analyze historical data and accurately predict the parking lot's revenue in a certain period of time in the future, so as to provide an important reference for the commercial layout of the parking lot. Summary of the invention

[0006] In order to solve the above-mentioned technical problem of parking lot revenue prediction, the present invention provides a parking lot revenue prediction method. The following technical solution is adopted:

[0007] A parking lot revenue prediction method comprises the following steps:

[0008] Step 1: Collect historical parking data of a set historical time period of the parking lot. The historical parking data includes multiple operating days in the historical time period. The operating days are classified into different operating day types according to the day of the week. If the operating day is a statutory holiday, a statutory holiday type is added. If the operating day is a store anniversary event day, a store anniversary event day type is added.

[0009] Step 2, according to the classification results, calculate the average toll traffic volume and average toll duration of each operating day category by category;

[0010] Step 3: Build a parking lot revenue prediction model based on deep learning algorithm to predict the set time period. The decision tree of the parking lot revenue prediction model decomposes the time period to be predicted into multiple daily parking lot revenue predictions in units of days. The input item of the daily parking lot revenue prediction is the corresponding operation day type. All daily parking lot revenue prediction results within the time period are accumulated to obtain the parking lot revenue prediction result within the time period.

[0011] By adopting the above technical solution, we can distinguish the traditional method of linearly predicting parking lot revenue through historical data analysis, and analyze the parking lot historical parking data in detail to obtain different types of operating days such as weekdays, weekends, holidays, and store anniversary events. The average toll traffic volume and average toll duration of different types of operating days are used as revenue parameters for future revenue estimation.

[0012] In the process of predicting parking lot revenue, it is necessary to first determine a future set time period, such as one month, half a month, or a set interval of time. This time period can actually be obtained through the calendar and other means to obtain the parking day types contained therein;

[0013] Based on the deep learning algorithm, a parking lot revenue prediction model can be constructed to predict the future set time period. The core decision formula of the parking lot revenue prediction model decision tree specifically considers how many different operating day types are included in the set time period, and then based on the average toll traffic and average charging duration of different operating day types in the historical data, the corresponding average toll traffic and average charging duration of different operating day types in the future set time period can be calculated, thereby achieving a more accurate prediction of the parking lot revenue on the operating day. The parking lot revenue of all operating days in the future set time period can then be accumulated to obtain the parking lot revenue prediction result for the future set time period.

[0014] The prediction method can achieve accurate and specific time prediction for a set time period in the future for different parking lots. Parking lot managers can better plan parking resources, such as adjusting the allocation of parking spaces, expanding or reducing parking areas, thereby optimizing resource allocation and improving operational efficiency.

[0015] The forecast results can help the finance department to formulate budgets more accurately, conduct cost control and financial planning, and reduce operating risks. Through revenue forecasting, market trends and consumer behavior can be analyzed, and marketing and pricing strategies can be adjusted in a timely manner to attract more users.

[0016] Optionally, in step 1, the historical parking data refers to the operating data of the parking lot in operation within 30-90 days.

[0017] By adopting the above technical solution, the use of historical parking data generally refers to the operation data of the parking lot in operation within 30-90 days. The recent historical data is more meaningful for the prediction of future parking lot revenue.

[0018] Optionally, in step 1, the operating days are classified into Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, statutory holidays and store anniversary event days.

[0019] By adopting the above technical solution, in the actual operation of the parking lot, the actual difference between different operating days is large. Monday to Friday is generally called the normal state. At this time point, if it is some non-industrial areas or office clusters, the traffic flow of the parking lot is actually poor. The corresponding normal state is actually very different every day. For example, for some specific office points, there may be large enterprises that adopt staggered rest periods nearby, so the traffic flow from Monday to Friday varies greatly. Therefore, the normal state is also subdivided into Monday, Tuesday, Wednesday, Thursday, and Friday;

[0020] Saturdays and Sundays are generally referred to as weekends, but for the same reason as above, there may be large-scale enterprises in the region that only have one day off. In this case, they need to be treated differently. On statutory holidays and store anniversary days, the traffic volume will increase dramatically, so they need to be classified separately.

[0021] This more subdivided type of operating day actually does not have a large amount of data and does not burden the computer processing, but it is of great significance to the subsequent parking lot revenue forecast and can improve the universality and accuracy of the parking lot revenue forecasting method.

[0022] Optionally, in step 2, based on the classification results of the operating days, the average toll traffic flow and average toll duration from Monday to Sunday, and the average toll traffic flow and average toll duration on statutory holidays are calculated respectively.

[0023] By adopting the above technical solution, the average toll traffic and the average toll time are meaningful. For the parking lot income, some unpaid parking times are meaningless to deal with. Here, only the charged vehicles are counted as toll traffic for statistical toll time.

[0024] Optionally, the core formula of the decision tree for the parking lot revenue prediction model is as follows:

[0025]

[0026] Where R represents the parking lot revenue forecast result from day a to day b, P Pv is the predicted toll traffic volume on date i, P PTis the predicted charging duration on date i, and Cs is the parking fee standard of the parking lot during the time period from day a to day b.

[0027] By adopting the above technical solution, the time period from day a to day b is the future time period to be predicted. For date i in the future time period to be predicted, the corresponding predicted toll traffic volume, predicted toll duration and parking fee standard are the input items of the prediction model. The above predicted toll traffic volume and predicted toll duration are the average toll traffic volume and average toll duration of the corresponding operating day classification of the parking lot operating day on the corresponding date, thereby realizing rapid and intelligent prediction of parking lot revenue.

[0028] Optionally, the decision tree of the parking lot revenue prediction model adds a daily average temperature impact factor γ, and the daily average temperature impact factor γ is determined by the following formula:

[0029]

[0030] Where β1 is the temperature coefficient in the linear regression model, which indicates the average change in parking lot revenue when the average temperature changes by one unit;

[0031] T a is the average temperature of the corresponding operating day classification, R a It is the average revenue of the category on the corresponding operating day.

[0032] By adopting the above technical solutions, the impact of temperature on parking lot revenue mainly reflects changes in parking demand and parking experience;

[0033] This reflects the change in parking demand: in hot weather, people are more inclined to look for indoor parking lots to avoid the impact of high temperatures, which may lead to an increase in parking demand for indoor parking lots, thereby increasing revenue. On the contrary, in cold weather, outdoor parking lots may be more popular, affecting the revenue of indoor parking lots.

[0034] High or low temperatures will affect customers' parking experience; if a parking lot can provide a comfortable parking environment (such as air conditioning, sunshade facilities, etc.), it can improve customer satisfaction, thereby increasing the proportion of repeat customers, which is beneficial to increase revenue in the long run. On the contrary, if the parking environment is not good, it may lead to customer loss. Therefore, for a specific parking lot, changes in temperature will have an impact on revenue.

[0035] Since the environmental conditions of parking lots are basically determined, it is necessary to study the effect of temperature on the revenue prediction of parking lots.

[0036] Since the time period is set in the future, such as half a month to a month, the corresponding daily temperature can actually be obtained through the official weather forecast channel, and the average temperature or the highest and lowest temperatures can be obtained by calculation. The daily average temperature is used here to assist in predicting parking lot income, and weather factors such as humidity and rain actually have an impact on parking lot income. However, the prediction accuracy of future time periods is generally poor after a week, and the accuracy of the average temperature prediction is generally still in a relatively accurate range. Here, when predicting parking lot income, adding the daily average temperature impact factor γ can help achieve more accurate parking lot income prediction results.

[0037] To calculate the impact factor of daily average temperature on parking revenue, we can use the concept of an impact factor, which usually refers to the relative percentage change of one variable when another variable changes by one unit.

[0038] Optionally, the daily average temperature impact factor γ is determined by the following steps:

[0039] Step a, using a linear regression model to fit the data and obtain the temperature coefficient β1;

[0040] Step b, calculate the average temperature T a and average income R a ;

[0041] Step c, obtain β1, average temperature T a and average income R a Substitute the daily average temperature impact factor γ into the determination formula to calculate the daily average temperature impact factor γ for different operating day classifications.

[0042] By adopting the above technical solution, for example, if β1=50 (meaning that for every 1°C increase in temperature, the average income increases by 50 yuan) is obtained through linear regression analysis, the average temperature is 20°C, and the average income is 2,000 yuan, then the impact factor is calculated as follows:

[0043]

[0044] This means that for every unit change in the daily average temperature, the parking lot revenue changes by 5% on average. This impact factor can help us understand the importance of temperature on parking lot revenue. If the impact factor is close to 0, it means that temperature has little impact on revenue; if the impact factor is much greater than 0, it means that temperature is an important factor.

[0045] Optionally, in step c, the daily average temperature influencing factors γ of different operating day classifications correspond to Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, statutory holidays and store anniversary days, and are recorded as γ1, γ2, γ3, γ4, γ5, γ6, γ7, γ7, and γ7 respectively.

[0046] By adopting the above technical solution, by separately calculating the daily average temperature impact factor γ of different operating day classifications, the temperature impact factor γ can be more finely introduced into the parking lot revenue prediction model, making the parking lot revenue prediction more accurate.

[0047] Optionally, the historical daily average temperature and future daily average temperature data can be obtained by fetching the future weather forecast officially released by the China Meteorological Administration through the Internet.

[0048] Optionally, the decision tree formula of the parking lot income prediction model with the daily average temperature impact factor γ is as follows:

[0049]

[0050] By adopting the above technical solution, the Internet can crawl the future weather forecast officially released by the China Meteorological Administration for one month. The parking lot revenue prediction model with the daily average temperature impact factor γ can make the parking lot revenue prediction results more accurate in the short term, such as the next 30 days.

[0051] For longer-term parking lot revenue forecasts, the daily average temperature impact factor γ can be removed because the weather forecast part of the longer-term parking lot revenue forecast has weaker reference significance.

[0052] In summary, the present invention includes at least one of the following beneficial technical effects:

[0053] The present invention can provide a parking lot revenue prediction method, which analyzes the parking lot's historical parking data in detail to obtain different types of operating days such as weekdays, weekends, holidays, and store anniversary events, and uses the average toll vehicle flow and average toll duration of different types of operating days as revenue parameters for future revenue calculation;

[0054] Based on the deep learning algorithm, a parking lot revenue prediction model can be constructed to predict the future set time period. The core decision formula of the parking lot revenue prediction model decision tree specifically considers how many different operating day types are included in the set time period, and then based on the average toll traffic and average charging duration of different operating day types in the historical data, the corresponding average toll traffic and average charging duration of different operating day types in the future set time period can be calculated, thereby achieving a more accurate prediction of the parking lot revenue on the operating day. The parking lot revenue of all operating days in the future set time period can then be accumulated to obtain the parking lot revenue prediction result for the future set time period.

[0055] The prediction method can achieve accurate and specific time prediction for a set time period in the future for different parking lots. Parking lot managers can better plan parking resources, such as adjusting the allocation of parking spaces, expanding or reducing parking areas, thereby optimizing resource allocation and improving operational efficiency.

[0056] The forecast results can help the finance department to formulate budgets more accurately, conduct cost control and financial planning, and reduce operating risks. Through revenue forecasting, market trends and consumer behavior can be analyzed, and marketing and pricing strategies can be adjusted in a timely manner to attract more users. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of a parking lot revenue prediction method of the present invention; DETAILED DESCRIPTION

[0058] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0059] The embodiment of the present invention discloses a parking lot income prediction method.

[0060] Reference Figure 1 , a parking lot revenue prediction method, comprising the following steps:

[0061] Step 1: Collect historical parking data of a set historical time period of the parking lot. The historical parking data includes multiple operating days in the historical time period. The operating days are classified into different operating day types according to the day of the week. If the operating day is a statutory holiday, a statutory holiday type is added. If the operating day is a store anniversary event day, a store anniversary event day type is added.

[0062] Step 2, according to the classification results, calculate the average toll traffic volume and average toll duration of each operating day category by category;

[0063] Step 3: Build a parking lot revenue prediction model based on deep learning algorithm to predict the set time period. The decision tree of the parking lot revenue prediction model decomposes the time period to be predicted into multiple daily parking lot revenue predictions in units of days. The input item of the daily parking lot revenue prediction is the corresponding operation day type. All daily parking lot revenue prediction results within the time period are accumulated to obtain the parking lot revenue prediction result within the time period.

[0064] Different from the traditional method of linearly predicting parking lot revenue through historical data analysis, the historical parking data of the parking lot is analyzed in detail to obtain different types of operating days such as weekdays, weekends, holidays, and store anniversary events. The average toll traffic volume and average toll duration of different operating day types are used as revenue parameters for future revenue estimation.

[0065] In the process of predicting parking lot revenue, it is necessary to first determine a future set time period, such as one month, half a month, or a set interval of time. This time period can actually be obtained through the calendar and other means to obtain the parking day types contained therein;

[0066] Based on the deep learning algorithm, a parking lot revenue prediction model can be constructed to predict the future set time period. The core decision formula of the parking lot revenue prediction model decision tree specifically considers how many different operating day types are included in the set time period, and then based on the average toll traffic and average charging duration of different operating day types in the historical data, the corresponding average toll traffic and average charging duration of different operating day types in the future set time period can be calculated, thereby achieving a more accurate prediction of the parking lot revenue on the operating day. The parking lot revenue of all operating days in the future set time period can then be accumulated to obtain the parking lot revenue prediction result for the future set time period.

[0067] The prediction method can achieve accurate and specific time prediction for a set time period in the future for different parking lots. Parking lot managers can better plan parking resources, such as adjusting the allocation of parking spaces, expanding or reducing parking areas, thereby optimizing resource allocation and improving operational efficiency.

[0068] The forecast results can help the finance department to formulate budgets more accurately, conduct cost control and financial planning, and reduce operating risks. Through revenue forecasting, market trends and consumer behavior can be analyzed, and marketing and pricing strategies can be adjusted in a timely manner to attract more users.

[0069] In step 1, the historical parking data refers to the operation data of the parking lot in operation within 30-90 days.

[0070] The use of historical parking data generally refers to the operating data of parking lots in operation within 30-90 days. Recent historical data is more meaningful for future parking lot revenue forecasts.

[0071] In step 1, the operating days are classified into Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, statutory holidays and store anniversary event days.

[0072] In the actual operation of the parking lot, the actual difference between different operating days is large. Monday to Friday is generally called the normal state. At this time point, if it is some non-industrial areas or office clusters, the traffic flow of the parking lot is actually poor. The corresponding normal state is actually very different every day. For example, for some specific office points, there may be large enterprises that take staggered breaks nearby, so the traffic flow from Monday to Friday varies greatly. Therefore, the normal state is also subdivided into Monday, Tuesday, Wednesday, Thursday, and Friday;

[0073] Saturdays and Sundays are generally referred to as weekends, but for the same reason as above, there may be large-scale enterprises in the region that only have one day off. In this case, they need to be treated differently. On statutory holidays and store anniversary days, the traffic volume will increase dramatically, so they need to be classified separately.

[0074] This more subdivided type of operating day actually does not have a large amount of data and does not burden the computer processing, but it is of great significance to the subsequent parking lot revenue forecast and can improve the universality and accuracy of the parking lot revenue forecasting method.

[0075] In step 2, based on the classification results of the operating days, the average toll traffic flow and average toll duration from Monday to Sunday, and the average toll traffic flow and average toll duration on statutory holidays are calculated respectively.

[0076] The average paid traffic flow and the average charging time are meaningful. For the parking lot income, some unpaid parking time is meaningless to deal with. Here, only the charged vehicles are counted as paid traffic flow for the statistical charging time.

[0077] The core formula of the decision tree for the parking lot revenue prediction model is as follows:

[0078]

[0079] Where R represents the parking lot revenue forecast result from day a to day b, P Pv is the predicted toll traffic volume on date i, P PT is the predicted charging duration on date i, and Cs is the parking fee standard of the parking lot during the time period from day a to day b.

[0080] The time period from day a to day b is the future time period to be predicted. For date i within the future time period to be predicted, the corresponding predicted toll traffic volume, predicted toll duration and parking fee standard are the input items of the prediction model. The above predicted toll traffic volume and predicted toll duration are the average toll traffic volume and average toll duration of the corresponding operating day classification of the parking lot operating day on the corresponding date, realizing rapid and intelligent prediction of parking lot revenue.

[0081] The decision tree of the parking lot revenue prediction model adds the daily average temperature impact factor γ. The formula for determining the daily average temperature impact factor γ is as follows:

[0082]

[0083] Where β1 is the temperature coefficient in the linear regression model, which indicates the average change in parking lot revenue when the average temperature changes by one unit;

[0084] T a is the average temperature of the corresponding operating day classification, R a It is the average revenue of the category on the corresponding operating day.

[0085] The impact of temperature on parking revenue mainly reflects changes in parking demand and parking experience;

[0086] This reflects the change in parking demand: in hot weather, people are more inclined to look for indoor parking lots to avoid the impact of high temperatures, which may lead to an increase in parking demand for indoor parking lots, thereby increasing revenue. On the contrary, in cold weather, outdoor parking lots may be more popular, affecting the revenue of indoor parking lots.

[0087] High or low temperatures will affect customers' parking experience; if a parking lot can provide a comfortable parking environment (such as air conditioning, sunshade facilities, etc.), it can improve customer satisfaction, thereby increasing the proportion of repeat customers, which is beneficial to increase revenue in the long run. On the contrary, if the parking environment is not good, it may lead to customer loss. Therefore, for a specific parking lot, changes in temperature will have an impact on revenue.

[0088] Since the environmental conditions of parking lots are basically determined, it is necessary to study the effect of temperature on the revenue prediction of parking lots.

[0089] Since the time period is set in the future, such as half a month to a month, the corresponding daily temperature can actually be obtained through the official weather forecast channel, and the average temperature or the highest and lowest temperatures can be obtained by calculation. The daily average temperature is used here to assist in predicting parking lot income, and weather factors such as humidity and rain actually have an impact on parking lot income. However, the prediction accuracy of future time periods is generally poor after a week, and the accuracy of the average temperature prediction is generally still in a relatively accurate range. Here, when predicting parking lot income, adding the daily average temperature impact factor γ can help achieve more accurate parking lot income prediction results.

[0090] To calculate the impact factor of daily average temperature on parking revenue, we can use the concept of an impact factor, which usually refers to the relative percentage change of one variable when another variable changes by one unit.

[0091] The daily average temperature impact factor γ is determined by the following steps:

[0092] Step a, using a linear regression model to fit the data and obtain the temperature coefficient β1;

[0093] Step b, calculate the average temperature T a and average income R a ;

[0094] Step c, obtain β1, average temperature T a and average income R a Substitute the daily average temperature impact factor γ into the determination formula to calculate the daily average temperature impact factor γ for different operating day classifications.

[0095] For example, if the linear regression analysis yields β1 = 50 (meaning that for every 1°C increase in temperature, the average income increases by 50 yuan), the average temperature is 20°C, and the average income is 2,000 yuan, then the impact factor is calculated as follows:

[0096]

[0097] This means that for every unit change in the daily average temperature, the parking lot revenue changes by 5% on average. This impact factor can help us understand the importance of temperature on parking lot revenue. If the impact factor is close to 0, it means that temperature has little impact on revenue; if the impact factor is much greater than 0, it means that temperature is an important factor.

[0098] In step c, the daily average temperature influencing factors γ of different operating day classifications correspond to Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, statutory holidays and store celebration days respectively, and are recorded as γ1, γ2, γ3, γ4, γ5, γ6, γ7, γ7, and γ7.

[0099] By calculating the daily average temperature impact factor γ for different operating day classifications separately, the temperature impact factor γ can be more finely introduced into the parking lot revenue prediction model, making the parking lot revenue prediction more accurate.

[0100] The historical and future daily average temperature data are obtained by fetching the future weather forecast officially released by the China Meteorological Administration through the Internet.

[0101] The decision tree formula of the parking lot income prediction model with the daily average temperature impact factor γ is as follows:

[0102]

[0103] The future weather forecast officially released by the China Meteorological Administration can be obtained from the Internet for one month. The parking revenue prediction model with the daily average temperature impact factor γ is more accurate for the short term, such as the next 30 days.

[0104] For longer-term parking lot revenue forecasts, the daily average temperature impact factor γ can be removed because the weather forecast part of the longer-term parking lot revenue forecast has weaker reference significance.

[0105] The following is a specific example to illustrate the implementation principle of a parking lot revenue prediction method of the present invention:

[0106] For a parking lot, its historical data is analyzed, and the data of different operating days are obtained based on the historical data of the parking lot in the previous 60 days as shown in Table 1:

[0107] Table 1

[0108] Operation day type Average toll traffic Average charging time Monday 354 1.83 Tuesday 325 1.74 Wednesday 362 1.75 Thursday 291 1.72 Friday 350 1.71 Saturday 619 1.82 Sunday 389 1.87 Statutory Holidays 556 2.12

[0109] First, we need to make a revenue forecast for the next seven days, with a time period of September 26, 2024 to October 2, 2024. We need to make a forecast for the parking lot revenue before and during the National Day, so as to provide reference data for the operation management before the National Day. The comparison between the predicted parking lot revenue forecast and the actual revenue results is shown in Table 2:

[0110] Table 2

[0111]

[0112] It can be seen that the gap between the predicted income of 17,854 yuan and the actual income of 17,676 yuan is small, and the prediction result is accurate.

[0113] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A parking lot revenue prediction method, characterized in that: The following steps are involved: Step 1: Collect historical parking data of a set historical time period of the parking lot. The historical parking data includes multiple operating days in the historical time period. The operating days are classified into different operating day types according to the day of the week. If the operating day is a statutory holiday, a statutory holiday type is added. If the operating day is a store anniversary event day, a store anniversary event day type is added. Step 2, according to the classification results, calculate the average toll traffic volume and average toll duration of each operating day category by category; Step 3: Build a parking lot revenue prediction model based on deep learning algorithm to predict the set time period. The decision tree of the parking lot revenue prediction model decomposes the time period to be predicted into multiple daily parking lot revenue predictions in units of days. The input item of the daily parking lot revenue prediction is the corresponding operation day type. All daily parking lot revenue prediction results within the time period are accumulated to obtain the parking lot revenue prediction result within the time period.

2. A parking lot income prediction method according to claim 1, characterized in that: In step 1, the historical parking data refers to the operation data of the parking lot in operation within 30-90 days.

3. A parking lot income prediction method according to claim 2, characterized in that: In step 1, the operating days are classified into Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, statutory holidays and store anniversary event days.

4. A parking lot income prediction method according to claim 3, characterized in that: In step 2, based on the classification results of the operating days, the average toll traffic flow and average toll duration from Monday to Sunday, and the average toll traffic flow and average toll duration on statutory holidays are calculated respectively.

5. A parking lot income prediction method according to claim 4, characterized in that: In step 3, the core formula of the decision tree of the parking lot revenue prediction model is as follows: Where R represents the parking lot revenue forecast result from day a to day b, P Pv is the predicted toll traffic volume on date i, P PT is the predicted charging duration on date i, and Cs is the parking fee standard of the parking lot during the time period from day a to day b.

6. A parking lot income prediction method according to claim 5, characterized in that: The decision tree of the parking lot revenue prediction model adds the daily average temperature impact factor γ. The formula for determining the daily average temperature impact factor γ is as follows: Where β1 is the temperature coefficient in the linear regression model, which indicates the average change in parking lot revenue when the average temperature changes by one unit; T a is the average temperature of the corresponding operating day classification, R a It is the average revenue of the category on the corresponding operating day.

7. A parking lot income prediction method according to claim 6, characterized in that: The daily average temperature influencing factor γ is determined by the following steps: Step a, using a linear regression model to fit the data and obtain the temperature coefficient β1; Step b, calculate the average temperature T a and average income R a ; Step c, obtain β1, average temperature T a and average income R a Substitute the daily average temperature impact factor γ into the determination formula to calculate the daily average temperature impact factor γ for different operating day classifications.

8. A parking lot income prediction method according to claim 7, characterized in that: In step c, the daily average temperature influencing factors γ of different operating day classifications correspond to Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, statutory holidays and store celebration days respectively, and are recorded as γ1, γ2, γ3, γ4, γ5, γ6, γ7, γ7, and γ7.

9. A parking lot income prediction method according to claim 8, characterized in that: The historical and future daily average temperature data are obtained by fetching the future weather forecast officially released by the China Meteorological Administration through the Internet.

10. A parking lot income prediction method according to claim 9, characterized in that: The decision tree formula of the parking lot income prediction model with the daily average temperature impact factor γ is as follows:

Images