Method and device for predicting number of network orders

Abstract

The invention relates to the technical field of computers, and discloses a method and device for predicting the number of network orders. The method comprises the following steps: predicting the number of network orders per day in future M days through a prediction model according to the number of network orders per day in historical N days; according to the order information of the plurality of network orders on the jth day in the historical N days, correcting the order placing time of the plurality of network orders, and according to the corrected order placing time of the plurality of network orders, determining the network order distribution condition of each hour in at least one hour on the jth day; and for the ith day in the future M days, according to the network order distribution condition of each hour in the at least one hour of the jth day and the network order quantity of the ith day, determining the network order quantity of each hour in the at least one hour of the ith day. Therefore, by correcting the order placing time of the plurality of network orders, the accuracy of the prediction method can be improved while the real demand of the user can be reflected.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method and device for predicting the quantity of online orders. Background Art

[0002] As the business grows, the fluctuation of the number of online orders has brought about a series of problems such as recruitment and staffing. Since the number of online orders of stores will fluctuate with the trend of time during business hours, the problem of store staffing will arise. Formulating corresponding scheduling strategies and transportation strategies based on the prediction results of the number of online orders is an important means at present.

[0003] The existing method for predicting the number of orders is to treat the number of online orders per hour on a certain historical day as a time series, and predict the number of online orders per hour on a certain future day according to a common time series prediction algorithm, such as an autoregressive integrated moving average model (ARIMA) or an exponential smoothing method.

[0004] However, the number of online orders predicted by the above method is not accurate enough. Therefore, how to accurately predict the number of online orders of stores is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The embodiment of the present invention provides a method and device for predicting the number of online orders, which are used to accurately predict the number of online orders of a store per hour in at least one hour in the future day.

[0006] In a first aspect, an embodiment of the present invention provides a method for predicting the number of online orders, the method comprising: predicting the number of online orders for each day in the future M days through a prediction model based on the number of online orders for each day in the history of N days, where N and M are both integers greater than or equal to 1; correcting the order time of multiple online orders on the jth day in the history of N days, and determining the distribution of online orders for each hour in at least one hour of the jth day based on the corrected order time of the multiple online orders; j is an integer greater than or equal to 1 and less than or equal to N; for the i-th day in the future M days, determining the number of online orders for each hour in at least one hour of the i-th day based on the distribution of online orders for each hour in at least one hour of the j-th day and the number of online orders on the i-th day; i is an integer greater than or equal to 1 and less than or equal to M.

[0007] By adopting the above method, the order time of historical online orders may not accurately reflect the real needs of users due to store order restrictions and other reasons. Therefore, by correcting the order time of multiple historical online orders, and determining the distribution of online orders for each hour in at least one hour of the jth day (referred to as hourly distribution) based on the corrected order time of multiple historical online orders, the determined hourly distribution can be made more reasonable and accurate. Furthermore, the number of online orders for each day in the next M days is predicted by the prediction model. For the i-th day in the next M days, the number of online orders per hour in at least one hour of the i-th day is determined based on the number of online orders on the i-th day and the hourly distribution. Compared with the method of predicting the number of hourly online orders based on the number of hourly online orders in the prior art, the accuracy of the predicted number of online orders per hour in at least one hour of the next day can be effectively improved, and the robustness of the prediction system can be improved.

[0008] Optionally, the order information of each online order among the multiple online orders includes at least one of the following: first information, the first information is used to indicate whether the online order is an instant order, and the order time of the instant order is the same as the user's expected delivery time; second information, the second information is used to indicate whether the online order has an order limit mark. When the online order has an order limit mark, it means that the online order is an online order that is delayed in submission due to the occurrence of an order limit; third information, the third information is used to indicate whether the online order is delayed in delivery due to capacity constraints.

[0009] In this way, based on the order information of each online order among the multiple online orders, a judgment basis is provided for the subsequent correction of the order placement time of the online orders.

[0010] Optionally, based on the order information of multiple online orders on the jth day in the history of N days, the order times of the multiple online orders are corrected, including: when the online order is an instant order and does not have a limit order mark, the order time of the online order is corrected to the user's expected delivery time; or, when the online order is an instant order and has a limit order mark, the order time of the online order is corrected to the user's expected order time; or, when the online order is a non-instant order, the order time of the online order is corrected to the user's expected delivery time.

[0011] Optionally, when the online order is a non-instant order, the order time of the online order is corrected to the user's expected delivery time, including: when the online order is a non-instant order and the order time of the online order is within a first time range, the order time of the online order is corrected to the user's expected delivery time; or, when the online order is a non-instant order and the order time of the online order is within a second time range, if the time interval between the user's expected delivery time and the order time is greater than or equal to a first time threshold, the order time is corrected to the user's expected delivery time.

[0012] Optionally, when the online order is a non-instant order and the order time of the online order is within a second time range, if the time interval between the user's expected delivery time and the order time is less than a first time threshold, and the delay time for delayed delivery due to capacity limitations is greater than or equal to a second time threshold, the order time is corrected to a time after the order time is shifted backward by a third time threshold; or, if the time interval between the user's expected delivery time and the order time is less than the first time threshold, and the delay time for delayed delivery due to capacity limitations is less than the second time threshold, there is no need to modify the order time.

[0013] By adopting the above method, by correcting the order time of the online order, the corrected online order can reflect the real needs of the user, and the distribution of online orders in each hour of at least one hour of the jth day in the future M days is determined based on the corrected online order, so as to ensure the accuracy of the subsequent prediction of the number of online orders in each hour of at least one hour of the jth day in the future M days.

[0014] Optionally, the types of the i-th day and the j-th day are the same, and the types of the i-th day and the j-th day are working days, weekends, or holidays.

[0015] Optionally, the prediction model is trained based on historical data of the store, and the historical data includes online order data of the store within one day and at least one of weather characteristics, activity information characteristics and holiday characteristics of that day.

[0016] In this way, adding weather features, activity information features, and holiday features to the features can enrich the diversity of features and thus improve the accuracy of model predictions.

[0017] In a second aspect, an embodiment of the present invention provides a network order quantity prediction device, comprising: an acquisition module, used to predict the number of network orders per day in the future M days through a prediction model based on the number of network orders per day in the history of N days, where N and M are both integers greater than or equal to 1; a correction module, used to correct the order time of multiple network orders on the jth day in the history of N days according to the order information of the multiple network orders, and determine the distribution of network orders for each hour in at least one hour of the jth day according to the corrected order time of the multiple network orders; j is an integer greater than or equal to 1 and less than or equal to N; a processing module, used to determine the number of network orders for each hour in at least one hour of the i-th day according to the distribution of network orders for each hour in at least one hour of the j-th day and the number of network orders on the i-th day for the i-th day in the future M days; i is an integer greater than or equal to 1 and less than or equal to M.

[0018] Optionally, the order information of each online order among the multiple online orders includes at least one of the following: first information, the first information is used to indicate whether the online order is an instant order, and the order time of the instant order is the same as the user's expected delivery time; second information, the second information is used to indicate whether the online order has an order limit mark. When the online order has an order limit mark, it means that the online order is an online order that is delayed in submission due to the occurrence of an order limit; third information, the third information is used to indicate whether the online order is delayed in delivery due to capacity constraints.

[0019] Optionally, the correction module is specifically used to: when the online order is an instant order and does not have a limit order mark, correct the order time of the online order to the user's expected delivery time; or, when the online order is an instant order and has a limit order mark, correct the order time of the online order to the user's expected order time; or, when the online order is a non-instant order, correct the order time of the online order to the user's expected delivery time.

[0020] Optionally, the correction module is specifically used to: when the online order is a non-instant order and the order time of the online order is within a first time range, correct the order time of the online order to the user's expected delivery time; or, when the online order is a non-instant order and the order time of the online order is within a second time range, if the time interval between the user's expected delivery time and the order time is greater than or equal to a first time threshold, correct the order time to the user's expected delivery time.

[0021] Optionally, the correction module is also used for: when the online order is a non-instant order and the order time of the online order is within a second time range, if the time interval between the user's expected delivery time and the order time is less than a first time threshold, and the delay time for delayed delivery due to capacity limitations is greater than or equal to a second time threshold, then the order time is corrected to a time after the order time is shifted backward by a third time threshold; or, if the time interval between the user's expected delivery time and the order time is less than the first time threshold, and the delay time for delayed delivery due to capacity limitations is less than the second time threshold, then there is no need to modify the order time.

[0022] Optionally, the types of the i-th day and the j-th day are the same, and the types of the i-th day and the j-th day are working days, weekends, or holidays.

[0023] Optionally, the acquisition module is also used to: obtain the prediction model based on training of historical data of the store, wherein the historical data includes online order data of the store within a day, and also includes at least one of weather characteristics, activity information characteristics and holiday characteristics of that day.

[0024] In a third aspect, an embodiment of the present invention provides a network order quantity prediction device, which includes: a memory for storing a computer program; and a processor for executing the computer program stored in the memory and executing the method described in the first aspect according to the obtained program.

[0025] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when a computer reads and executes the computer program, the method described in the first aspect above is executed. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0027] Figure 1 A schematic diagram of a user's order placement time shift caused by order restrictions provided in an embodiment of the present invention;

[0028] Figure 2 A flowchart corresponding to the method for predicting the quantity of online orders provided by an embodiment of the present invention;

[0029] Figure 3 A schematic diagram of a limited order-marked network order provided by an embodiment of the present invention;

[0030] Figure 4 A flowchart of a method for predicting the quantity of online orders provided by an embodiment of the present invention;

[0031] Figure 5 A schematic diagram of a network order quantity prediction device provided by an embodiment of the present invention;

[0032] Figure 6 A schematic diagram of the structure of a device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention. In the embodiments of the present invention, multiple refers to two or more. The words "first", "second", etc. are only used to distinguish the purpose of description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0034] Since the number of online orders of a store will fluctuate with the time trend during business hours, the problem of manpower allocation in the forward warehouse will arise. For example, more manpower can be arranged during periods with a large number of online orders, and part-time staff can be added; full-time staff can be arranged during periods with a small number of online orders. At present, the existing solution is to use the prediction model to predict the number of online orders per hour in the next M days based on the number of online orders per hour in the past N days, where N and M are both integers greater than or equal to 1.

[0035] However, traditional prediction methods only consider the information of the sequence itself, as well as some of its own periodic information, and directly predict the number of hourly orders in the next M days based on the number of historical online orders per hour. Other factors that cause changes in the number of online orders, such as the impact of weather, event information, holidays and other variables on the number of online orders, are not fully considered. In actual applications, factors such as order limits may occur, causing the user's order time to shift backwards. At this time, online orders cannot reflect the user's real needs, thereby reducing the accuracy of the model's prediction results. For example, see Figure 1 As shown in the figure, the store's daily user demand peak is 8-10 o'clock. Due to the order limit, the user order peak shifts to 12 o'clock. It can be seen that the order limit not only affects the user experience, but also affects the stability of the original network order data and subsequent use.

[0036] Based on this, an embodiment of the present application provides a method for predicting the number of online orders, which is used to accurately predict the number of online orders of a store per hour in at least one hour in the future day.

[0037] The method provided in the embodiments of the present application is described in detail below in conjunction with specific embodiments.

[0038] Figure 2 This is a flowchart corresponding to the network order quantity prediction method provided by the embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:

[0039] Step 201, based on the number of online orders per day in the past N days, the number of online orders in the next M days is predicted by a prediction model.

[0040] Here, predicting the number of online orders in the next M days through the prediction model is the basis for subsequently calculating the number of online orders in each hour of at least one hour of the i-th day in the next M days. The values ​​of N and M can be set according to actual needs, for example, N is equal to 30 and M is equal to 14.

[0041] For example, there are many prediction models for predicting the number of online orders within the next M days. Since the traditional time series prediction model is sensitive to outliers, if the prediction value sequence is mixed with outliers, the prediction value will be inaccurate and the prediction stability will be poor. Therefore, the embodiment of the present invention adopts a tree model to predict the number of online orders within the next M days. The tree model includes: decision tree (DT), random forest (RF) and boosting regression tree (BRT).

[0042] The generation of the prediction model is introduced below.

[0043] In an embodiment of the present invention, the prediction model can be trained based on the historical data of the store. The historical data of the store may include the online order data of the store on a daily basis over multiple historical days, and may also include at least one of the weather characteristics, activity information characteristics, and holiday characteristics of each day. Considering that the sales volume of the store may be affected by weather, holidays, and activity information, weather characteristics, activity information characteristics, and holiday characteristics may be introduced when training the prediction model.

[0044] The store here refers to the store that meets the preset conditions. For example, if the store has been in normal business status for one or more days in the past and will be in normal business status in the next M days, then it can be considered that the store meets the prediction conditions. In this way, it is easy to provide reasonable training data for the prediction model.

[0045] As shown in Table 1, weather characteristics include: maximum temperature, minimum temperature, temperature level, temperature level aggregate quantity, precipitation, precipitation level, wind force and other characteristics.

[0046] The activity information features include the number of single product activity SKUs, the proportion of single product activity SKUs, the average discount of single product activities, the number of special price single product activity SKUs, the proportion of special price single product activity SKUs, the average discount of special price single product activities, the number of total price activity SKUs, the proportion of total price activity SKUs, and the average discount of total price activities.

[0047] For example, taking the calculation of the number of activity sku for a single product as an example, firstly, it is necessary to clarify all possible attributes of the single product. For example, the single product A involves three activities D, E, and F. The number of activity sku for the single product is the order quantity of the three combinations of AD, AE, and AF. The proportion of activity sku for the single product is the number of activity sku for the single product / the total order quantity of single product A.

[0048] Holiday features include working days and rest days. When holiday_level is 0, it means there is no holiday impact. The difference in order quantity between working days is small, so there is no need to distinguish between the days of the week. When holiday_level is 1, it means rest days, which include special holidays. The importance of special holidays in a year is distinguished according to the holiday level. For example, the holiday level of Spring Festival is 4, the holiday level of National Day is 3, the holiday level of Labor Day is 2, the holiday level of New Year's Day, Qingming Festival, Dragon Boat Festival, and Mid-Autumn Festival is 1, and the holiday level of the remaining holidays is 0.

[0049] Table 1. Example of characteristics of historical online orders

[0050]

[0051]

[0052] In addition, the store's historical data also includes other possible features, which are not specifically limited. It is understood that if there are outliers in the acquired features, the outliers are removed and the missing values ​​are filled with the average number of orders in the past week. After the prediction model is trained, new features can be obtained based on new historical network orders to update the model, thereby improving the accuracy of the model prediction.

[0053] Step 202, based on the order information of multiple online orders on the jth day in the history of N days, the order time of multiple online orders is corrected, and based on the corrected order time of the multiple online orders, the distribution of online orders for each hour in at least one hour on the jth day is determined.

[0054] In an embodiment of the present invention, based on order information of multiple online orders on the jth day in historical N days, the order information of each online order in the multiple online orders includes at least one of the following, such as first information, second information and third information.

[0055] The first information is used to indicate whether the online order is an instant order. If the user chooses immediate delivery when placing an order, and the user's order time is the same as the user's expected delivery time, then the online order is an instant order; if the user chooses scheduled delivery when placing an order, that is, the user's order time is not the same as the user's expected delivery time, then the online order is a non-timely order.

[0056] Second information: used to indicate whether the online order has a limit order mark. If the online order is delayed due to a limit order, the online order has a limit order mark. Figure 3 As shown, the user wants to place an order at 9 o'clock. The order can be an immediate order or a non-immediate order. Since the store has set an order limit from 8 to 10 o'clock, the user cannot place the order at 9 o'clock and the online order is submitted after 10 o'clock. At this time, the order has a limit order mark.

[0057] The third information is used to indicate whether the online order is delayed due to capacity constraints. For example, the number of orders in the current period is 60, and the current capacity only supports the delivery of 50 orders. At this time, the excess 10 orders are delayed due to capacity constraints.

[0058] Exemplarily, the order information of each online order among the multiple online orders only includes the first information. At this time, the online orders are divided into two categories: instant orders and non-instant orders, and the order time of the two categories of orders is corrected.

[0059] When the online order is not an immediate order, the order time of the online order is corrected to the user's expected delivery time; when the online order is not an immediate order, if the order time of the online order is within the first time range, the order time of the online order is corrected to the user's expected delivery time; if the order time of the online order is within the second time range, there is no need to modify the order time. For example, the first time range is set to 0-6 o'clock, and the second time range is set to 6-23 o'clock. If the user places an order at 3 o'clock and the expected delivery time is 10 o'clock, it means that the user's actual demand is at 10 o'clock, and the user's order time at 3 o'clock is corrected to 10 o'clock; if the user places an order at 11 o'clock and the user's expected delivery time is 12 o'clock, there is no need to modify the order time.

[0060] Exemplarily, the order information of each online order in the multiple online orders includes first information, second information and third information, and the online orders can be divided into 4 categories, such as the first category, the second category, the third category and the fourth category.

[0061] Category 1: The online order is an instant order and does not have a limit order mark. The order time of the online order is corrected to the user's expected delivery time. Since the order time of the instant order is the same as the user's expected delivery time, the time of the online order does not need to be modified.

[0062] Category 2: Online orders are instant orders with order limit marks. The order time of online orders is corrected to the user's expected order time. For example, if a user wants to place an order at 9 o'clock, but the store sets an order limit from 8 to 10 o'clock, the user cannot place the order at 9 o'clock and the online order is submitted after 10 o'clock. The order time of the online order is corrected from 10 o'clock to 9 o'clock.

[0063] Category 3: When the online order is a non-instant order and the order time is within the first time range, the order time of the online order is corrected to the user's expected delivery time. For example, if the first time range is set to 0-6 o'clock, if the user places an order at 3 o'clock and the expected delivery time is 10 o'clock, the user's order time is corrected from 3 o'clock to 10 o'clock.

[0064] Category 4: When the online order is a non-instant order and the order time is within the second time range, if the time interval between the user's expected delivery time and the order time is greater than or equal to the first time threshold, the order time will be corrected to the user's expected delivery time; if the time interval between the user's expected delivery time and the order time is less than the first time threshold, and the delay time for delayed delivery due to capacity constraints is greater than or equal to the second time threshold, the order time will be corrected to the time after the order time is shifted backward by the third time threshold; if the time interval between the user's expected delivery time and the order time is less than the first time threshold, and the delay time for delayed delivery due to capacity constraints is less than the second time threshold, there is no need to modify the order time. For example, based on the second time range set above, the first time threshold is set to 3 hours, the second time threshold is set to 30 minutes, and the third time threshold is set to 1 hour. The user's order time is 10 o'clock. If the user expects delivery time at 14 o'clock, the time interval between the user's expected delivery time and the order time is 4 hours, which is greater than the first time interval of 3 hours. The order time is corrected from 10 o'clock to the user's expected delivery time of 14 o'clock; if the user expects delivery time at 12 o'clock, the time interval between the user's expected delivery time and the order time is 2 hours, which is less than the first time interval of 3 hours. If the delay time for delayed delivery is 40 minutes due to the order collection and transportation capacity limitation, which is greater than the second time threshold of 30 minutes, the order time is corrected from 10 o'clock to 11 o'clock, the time after the order time is shifted back by the third time threshold. If the delay time for delayed delivery is 20 minutes due to the order collection and transportation capacity limitation, which is less than the second time threshold of 30 minutes, there is no need to modify the order time.

[0065] Based on the corrected online orders, the number of orders per hour in the corrected online orders on the jth day and the total number of orders on the jth day are selected to determine the distribution of online orders in each hour in at least one hour on the jth day (which may be referred to as hourly distribution), wherein the data of the jth day is selected based on the type of the ith day, that is, the types of the ith day and the jth day are the same. The hourly distribution of the jth day can be the ratio of the number of orders per hour on the jth day in the historical N days to the total number of orders on the jth day, or it can be the average ratio of the number of orders per hour on the same type of days in the historical N days to the total number of orders per day on the same type of days in the N days, and there is no specific limitation. Taking the hourly distribution of the jth day as the ratio of the number of orders per hour on the jth day in the N historical days to the total number of orders on the jth day as an example, assuming that the type of the i-th day is a working day, Monday, select the online order data of the jth day of the same type as the i-th day in the N historical days, the j-th day is the previous Monday, and the hourly distribution of the j-th day is determined based on the ratio of the number of orders per hour on the j-th day to the total number of orders on the j-th day; assuming that the j-th day is the Dragon Boat Festival, select the online order data of the j-th day of the same type as the i-th day in the historical days, the j-th day is the Dragon Boat Festival of the previous year, and the hourly distribution of the j-th day is determined based on the ratio of the number of orders per hour on the j-th day to the total number of orders on the j-th day.

[0066] In practical applications, when selecting online order data for the jth day in history, there may be a lack of holiday information. In this case, the hourly online order distribution of holidays of the same level and stage at the same store level can be obtained to replace the hourly distribution of the jth day. Take the Dragon Boat Festival as an example. As shown in Table 1, the holiday level of the Dragon Boat Festival is 1. The holidays of level 1 include New Year's Day, Qingming Festival and Mid-Autumn Festival. The hourly distribution of holidays of the same level can be used as a replacement. When the data of holidays of the same level are missing, the store information is used to obtain the hourly online order distribution. The store information includes the city, province and region where the store is located. Select the same holiday at the same city level or the same or next level holiday at the city level. If the corresponding data cannot be obtained at the city level, obtain the same holiday at the province level or the same or next level holiday at the province level. If the corresponding data cannot be obtained at the province level, obtain the same holiday at the regional level or the same or next level holiday at the regional level.

[0067] Step 203, for the i-th day in the future M days, determine the number of online orders in each hour of at least one hour of the i-th day based on the distribution of online orders in each hour of at least one hour of the j-th day and the number of online orders on the i-th day.

[0068] In the embodiment of the present invention, the number of online orders in each hour in at least one hour of the i-th day is obtained based on the number of online orders on the i-th day and the hourly distribution on the j-th day, and the calculation formula is as follows:

[0069]

[0070] in, represents the number of online orders in the ath hour of the i-th day; Z i : The number of online orders on the i-th day predicted by the prediction model; It represents the network order distribution of hour a in the network order distribution of at least one hour in day j. For example, the prediction model predicts the network order data for the next 14 days, taking i as 1 as an example (the prediction results of the network order quantity for each hour of the remaining 13 days can be used as a reference), and the i-th day is Saturday, taking a as 10 o'clock as an example, represents the number of online orders at the 10th hour of the first day in the next 14 days. At this time, the distribution of online orders for each hour in at least one hour of last Saturday after the correction of historical online orders is obtained. j is set to 6. It represents the distribution of online orders in the 10th hour of the revised online orders last Saturday. Based on the above content, the number of online orders for each hour on the i-th day can be predicted.

[0071] Using the above method, due to store order restrictions and other reasons, the order time of historical online orders may not accurately reflect the user's real needs, such as Figure 4 As shown: by correcting the order time in multiple historical online order data, and determining the hourly distribution of online orders on the jth day based on the corrected order time of multiple historical online orders, the determined hourly distribution can be made more reasonable and accurate. Furthermore, the number of online orders per day in the next M days is predicted by the prediction model. For the i-th day in the next M days, the number of online orders per hour in at least one hour of the i-th day is determined based on the number of online orders and the hourly distribution on the i-th day. Compared with the method of predicting the number of hourly online orders based on the number of hourly online orders in the prior art, the accuracy of the predicted number of online orders per hour of the store in at least one hour of the next day can be effectively improved, and the robustness of the prediction system can be improved.

[0072] Based on the above prediction method, the number of online orders per hour in at least one hour in the next M days is obtained. The difference between the predicted number of online orders per hour in at least one hour in the next M days and the actual value can be analyzed to evaluate the accuracy of the prediction method. There are many methods for evaluating the accuracy of the prediction method, such as evaluation method 1 and evaluation method 2.

[0073] Evaluation method one: Draw a trend curve using the predicted results and true values ​​of the prediction method. That is to say, draw a trend curve of the predicted number of online orders per hour and the actual number of orders in at least one hour of the M days in chronological order to observe the degree of fit between the trend of the predicted number of online orders per hour and the actual number of online orders per hour. If the degree of fit is high, it means that the prediction method of the embodiment of the present invention is highly accurate; otherwise, it means that the accuracy is not high.

[0074] Evaluation method 2: The accuracy of the prediction method of the embodiment of the present invention is evaluated by two indicators: mean absolute percentage error (MAPE) and mean absolute error (MAE). The MAPE indicator is usually used to evaluate the deviation rate before the predicted value, and the MAE indicator is used to evaluate the absolute deviation between the predicted value and the actual value. The accuracy of the prediction method of the embodiment of the present invention can be measured based on the two indicators of MAPE and MAE.

[0075] The above embodiment can be applied to intelligent scheduling and recruitment business. Based on the prediction method of the embodiment of the present invention, the number of online orders per hour in at least one hour in the next M days is predicted to specify the corresponding scheduling strategy and recruitment strategy, thereby improving the utilization rate of human resources and reducing the company's costs.

[0076] Based on the same concept, the embodiment of the present invention also provides a network order quantity prediction device. Figure 5 A schematic diagram of a network order quantity prediction device provided by an embodiment of the present invention, such as Figure 5 As shown, including:

[0077] The acquisition module 501 is used to predict the number of online orders per day in the next M days based on the number of online orders per day in the past N days through a prediction model, where N and M are both integers greater than or equal to 1;

[0078] The correction module 502 is used to correct the order placement time of the multiple online orders on the jth day in the history of N days according to the order information of the multiple online orders, and determine the distribution of the online orders in each hour of at least one hour of the jth day according to the corrected order placement time of the multiple online orders; j is an integer greater than or equal to 1 and less than or equal to N;

[0079] Processing module 503 is used to determine, for the i-th day in the future M days, the number of online orders in each hour of at least one hour of the i-th day based on the distribution of online orders in each hour of at least one hour of the j-th day and the number of online orders on the i-th day; i is an integer greater than or equal to 1 and less than or equal to M.

[0080] Optionally, the order information of each online order among the multiple online orders includes at least one of the following: first information, the first information is used to indicate whether the online order is an instant order, and the order time of the instant order is the same as the user's expected delivery time; second information, the second information is used to indicate whether the online order has an order limit mark. When the online order has an order limit mark, it means that the online order is an online order that is delayed in submission due to the occurrence of an order limit; third information, the third information is used to indicate whether the online order is delayed in delivery due to capacity constraints.

[0081] Optionally, the correction module 502 is used to correct the order time of multiple online orders based on the order information of the jth day in the historical N days, including: when the online order is an instant order and does not have a limit order mark, correcting the order time of the online order to the user's expected delivery time; or, when the online order is an instant order and has a limit order mark, correcting the order time of the online order to the user's expected order time; or, when the online order is a non-instant order, correcting the order time of the online order to the user's expected delivery time.

[0082] Optionally, the correction module 502 is also used to correct the order time of the online order to the user's expected delivery time when the online order is a non-instant order, including: when the online order is a non-instant order and the order time of the online order is within a first time range, correcting the order time of the online order to the user's expected delivery time; or, when the online order is a non-instant order and the order time of the online order is within a second time range, if the time interval between the user's expected delivery time and the order time is greater than or equal to a first time threshold, correcting the order time to the user's expected delivery time.

[0083] Optionally, the correction module 502 is also used to, when the online order is a non-instant order and the order time of the online order is within a second time range, if the time interval between the user's expected delivery time and the order time is less than a first time threshold, and the delay time for delayed delivery due to capacity limitations is greater than or equal to a second time threshold, then correct the order time to a time after the order time is shifted backward by a third time threshold; or, if the time interval between the user's expected delivery time and the order time is less than the first time threshold, and the delay time for delayed delivery due to capacity limitations is less than the second time threshold, then there is no need to modify the order time.

[0084] Optionally, the types of the i-th day and the j-th day are the same, and the types of the i-th day and the j-th day are working days, weekends, or holidays.

[0085] Optionally, the acquisition module 501 obtains the prediction model based on training of historical data of the store, where the historical data includes online order data of the store within a day and at least one of weather characteristics, activity information characteristics and holiday characteristics of that day.

[0086] Based on the same concept, an embodiment of the present invention provides a computing device, Figure 6 FIG. 6 is a schematic diagram of a structure of a device 6000 provided in an embodiment of the present invention, and the device is used to implement the function of the first communication device in the above embodiment. Figure 6 As shown, the device 6000 includes at least one processor 601 and a memory 602 connected to the at least one processor 601. The specific connection medium between the processor 601 and the memory 602 is not limited in the embodiment of the present application. Figure 6 For example, the processor 601 and the memory 602 are connected via a bus. The bus can be divided into an address bus, a data bus, a control bus, etc. In an embodiment of the present invention, the memory 602 stores instructions that can be executed by at least one processor 601, and the at least one processor 601 can implement the steps of the above-mentioned method for predicting the number of network orders by executing the instructions stored in the memory 602.

[0087] The processor 601 is the control center of the computer device, and can use various interfaces and lines to connect various parts of the computer device, and perform resource settings by running or executing instructions stored in the memory 602 and calling data stored in the memory 602. Optionally, the processor 601 may include one or more processing units, and the processor 601 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 601. In some embodiments, the processor 601 and the memory 602 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on independent chips.

[0088] Processor 601 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware processor for execution, or can be executed by a combination of hardware and software modules in the processor.

[0089] The memory 602 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 602 may include at least one type of storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory, a random access memory (RAM), a static random access memory (SRAM), a programmable read-only memory (PROM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic memory, a disk, an optical disk, and the like. The memory 602 is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 602 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0090] Based on the same technical concept, an embodiment of the present invention provides a computer-readable storage medium, in which a computer program is stored, and the computer program is executed by a processor to implement the above-mentioned network order quantity prediction method.

[0091] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0092] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0093] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0094] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0095] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for predicting the number of online orders, characterized in that: The method comprises: Based on the number of online orders per day in the past N days, the forecasting model is used to predict the number of online orders per day in the next M days, where N and M are both integers greater than or equal to 1; According to order information of multiple online orders on the jth day in the history of N days, the order placement time of the multiple online orders is corrected, and according to the corrected order placement time of the multiple online orders, the distribution of online orders in each hour of at least one hour of the jth day is determined; j is an integer greater than or equal to 1 and less than or equal to N; For the i-th day in the future M days, determine the number of online orders in each hour of at least one hour of the i-th day based on the distribution of online orders in each hour of at least one hour of the j-th day and the number of online orders on the i-th day; i is an integer greater than or equal to 1 and less than or equal to M.

2. The method according to claim 1, characterized in that The order information of each of the multiple online orders includes at least one of the following: first information, the first information being used to indicate whether the online order is an instant order, and the order time of the instant order is the same as the user's expected delivery time; second information, the second information is used to indicate whether the online order has an order limit mark, and when the online order has an order limit mark, it means that the online order is an online order that is delayed in submission due to an order limit; The third information is used to indicate whether the online order is delayed in delivery due to transportation capacity constraints.

3. The method according to claim 2, characterized in that According to order information of multiple online orders on the jth day in the history of N days, the order placement time of the multiple online orders is corrected, including: When the online order is an instant order and does not have a limited order mark, the order time of the online order is corrected to the user's expected delivery time; or, When the online order is an instant order and has a limit order mark, the order time of the online order is corrected to the user's expected order time; or, When the online order is a non-instant order, the order time of the online order is corrected to the user's expected delivery time.

4. The method according to claim 3, characterized in that When the online order is a non-instant order, the order time of the online order is corrected to the user's expected delivery time, including: When the online order is a non-instant order and the order time of the online order is within the first time range, the order time of the online order is corrected to the user's expected delivery time; or, When the online order is a non-instant order and the order time of the online order is within the second time range, if the time interval between the user's expected delivery time and the order time is greater than or equal to the first time threshold, the order time is corrected to the user's expected delivery time.

5. The method according to claim 4, characterized in that The method further comprises: When the online order is a non-instant order and the order time of the online order is within the second time range, if the time interval between the user's expected delivery time and the order time is less than the first time threshold, and the delay time for delayed delivery due to capacity constraints is greater than or equal to the second time threshold, the order time is corrected to the time after the order time is shifted backward by a third time threshold; or, if the time interval between the user's expected delivery time and the order time is less than the first time threshold, and the delay time for delayed delivery due to capacity constraints is less than the second time threshold, there is no need to modify the order time.

6. The method according to any one of claims 1 to 5, characterized in that The types of the i-th day and the j-th day are the same, and the types of the i-th day and the j-th day are working days, weekends, or holidays.

7. The method according to any one of claims 1 to 5, characterized in that The prediction model is trained based on the historical data of the store, where the historical data includes the online order data of the store within a day and at least one of the weather characteristics, activity information characteristics and holiday characteristics of the day.

8. A device for predicting the number of online orders, characterized in that: The device comprises: The acquisition module is used to predict the number of online orders per day in the next M days based on the number of online orders per day in the past N days through a prediction model, where N and M are both integers greater than or equal to 1; a correction module, configured to correct the order placement time of the plurality of online orders on the jth day in the history of N days according to the order information of the plurality of online orders, and determine the distribution of the online orders in each hour in at least one hour of the jth day according to the corrected order placement time of the plurality of online orders; j is an integer greater than or equal to 1 and less than or equal to N; A processing module is used to determine, for the i-th day in the future M days, the number of online orders in each hour of at least one hour of the i-th day based on the distribution of online orders in each hour of at least one hour of the j-th day and the number of online orders on the i-th day; i is an integer greater than or equal to 1 and less than or equal to M.

9. A device for predicting the number of online orders, characterized in that: include: A memory for storing program instructions; A processor is used to call the program instructions stored in the memory, and execute the method according to any one of claims 1 to 7 according to the obtained program instructions.

10. A computer-readable storage medium, characterized in that: The method comprises computer-readable instructions, and when a computer reads and executes the computer-readable instructions, the method according to any one of claims 1 to 7 is executed.

Images