Hotel guest room state and price prediction method based on big data

Abstract

The invention discloses a hotel guest room state and price prediction method based on big data. The method comprises the following steps: matching a target guest room meeting guest room requirements and reservation time for a user; obtaining the current price of the target guest room and the guest room remaining amount of the target hotel containing the target guest room; predicting the guest room reservation growth rate of the target hotel at the user reservation time according to the related event big data; calculating an empty room rate of the target hotel at the user reservation time according to the current room remaining amount of the target hotel and the predicted room reservation growth rate at the user reservation time; and according to the current price of the target guest room and the empty room rate of the target hotel in the user reservation time, predicting a target guest room prediction price of the user reservation time. According to the invention, the guest room reservation growth rate and the empty room rate of the hotel at the future moment are predicted through the related event big data, the guest room predicted price is given to assist the user and the merchant in bargaining, and the use satisfaction of the user and the accuracy and efficiency of hotel guest room bargaining can be improved.

Description

Technical Field

[0001] The present invention relates to the field of hotel reservation and big data technology, and in particular to a hotel room status and price prediction method based on big data. Background Art

[0002] With the development of tourism and the Internet industry, some companies that operate hotel booking agency services through websites have emerged in the market, such as Ctrip, eLong, Mango.com, etc. These websites can display hotel profiles, room status, room types, room prices and other information, and users can book hotel rooms online or by phone through website intermediaries.

[0003] However, these websites have a flaw, that is, users can only book hotel rooms from website intermediaries according to the prices displayed on the website, and cannot communicate and negotiate prices directly with one or several hotel merchants, resulting in users only passively accepting standardized services, with little choice and no bargaining. For users, users hope to communicate with multiple hotels and obtain preferential prices through a competitive bargaining model; for hotels, when there are vacant rooms, the hotel is willing to establish direct communication with interested users, and after bargaining, sell the vacant rooms at a preferential price. As a special commodity, hotel rooms have a strong timeliness of value. Once the room is not sold in time, its value will be zero after expiration. If the vacant rooms are sold at a low price in a non-public and bargaining manner, the hotel can still obtain good profits after spending little costs such as corresponding cleaning fees, water and electricity fees; and it will not cause the disclosure of the preferential price, thereby impacting its normal market price system; at the same time, users will benefit.

[0004] However, the current bargaining platform does not provide reference prices for users and merchants. On the one hand, users do not know how to bid during the bargaining process. If the price is too low, the merchant will not accept it. If the price is too high, it will cause losses to the user. In addition, since there is no reference price, users cannot judge whether the merchant's counter-offer is reasonable, and will not approve the merchant's counter-offer, resulting in poor user satisfaction. On the other hand, for merchants, they generally set prices and counter-offers based on the current vacancy rate (i.e. the proportion of vacant guest rooms). However, when the user's reservation time is relatively late (such as a time one or two months after the user's reservation), the merchant cannot estimate the future room demand, i.e. the vacancy rate. At this time, the merchant cannot give accurate pricing and counter-offers, resulting in poor accuracy and efficiency in bargaining between users and merchants. Summary of the invention

[0005] In view of the deficiencies of the above-mentioned prior art, the technical problem to be solved by the present invention is: how to provide a hotel room status and price prediction method based on big data, predict the hotel's room reservation growth rate and vacancy rate in the future through relevant event big data, and then give the predicted room price to assist users and merchants in bargaining, which can not only improve user satisfaction, but also improve the accuracy and efficiency of hotel room bargaining.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] Hotel room status and price prediction methods based on big data include:

[0008] S1: Obtain the user's room requirements and booking time;

[0009] S2: Match the target room that meets the user's room requirements and booking time; obtain the current price of the target room and the remaining number of rooms in the target hotel that includes the target room;

[0010] S3: Obtain relevant event big data associated with the user's booking time, and predict the room booking growth rate of the target hotel at the user's booking time based on the relevant event big data;

[0011] S4: Calculate the vacancy rate of the target hotel at the user's booking time according to the current remaining number of rooms of the target hotel and the predicted room booking growth rate at the user's booking time;

[0012] S5: Predict the target room price at the user's booking time based on the current price of the target room and the vacancy rate of the target hotel at the user's booking time.

[0013] Preferably, the method further comprises step S6;

[0014] S6: When the user's reservation time includes multiple dates, execute steps S2 to S5 for each date to predict and obtain the target room predicted price for each date; based on the target room predicted price for each date, generate the target room predicted price change trend within the user's reservation time.

[0015] Preferably, in step S3, the relevant event big data includes holiday big data;

[0016] Forecast room booking growth rate by following these steps:

[0017] S301: The acquired holiday big data includes holiday types and holiday time periods, as well as historical room reservation big data of all hotels during holidays;

[0018] S302: Determine the overlapping date between the user's reservation time and the holiday time period;

[0019] S303: Obtain historical room reservation big data of the target hotel on the holiday type, and filter out the historical room reservation rate of the target hotel on the overlapping dates;

[0020] S304: Obtain historical room reservation big data of other hotels on this type of holiday, and filter out historical room reservation rates of other hotels on overlapping dates;

[0021] S305: Calculate the room reservation growth rate of the target hotel at the user's reservation time based on the historical room reservation rates of the target hotel and other hotels on overlapping dates.

[0022] Preferably, in step S3, the relevant event big data includes tourist population big data;

[0023] Forecast room booking growth rate by following these steps:

[0024] S311: The acquired tourist population big data includes tourist peak types and tourist peak time periods, as well as historical foreign population flow big data and historical room reservation big data of all hotels during tourist peaks;

[0025] S312: When the current date is a tourist peak period, obtain the inbound population flow of the current date;

[0026] S313: matching the historical foreign population flow rate with the foreign population flow rate on the current date from the tourist peak time period according to the historical foreign population flow rate big data;

[0027] S314: Calculate the interval N between the current date and the user's reservation time; push back N days the tourist peak date with the closest inflow of foreign people to the current date as the target tourist peak date corresponding to the user's reservation time;

[0028] S315: Obtain the historical room reservation rate of the target hotel on the target travel peak date corresponding to the user's reservation time;

[0029] S316: Obtain the historical room reservation rate of the target travel peak date corresponding to the reservation time of other hotel sub-users;

[0030] S317: Calculate the room reservation growth rate of the target hotel at the user's reservation time based on the historical room reservation rates of the target hotel and other hotels at the user's reservation time.

[0031] Preferably, in step S3, the relevant event big data includes aggregation activity big data;

[0032] Forecast room booking growth rate by following these steps:

[0033] S321: The acquired gathering activity big data includes the gathering activity type, gathering activity location, gathering activity time period and the number of people who have reported for the gathering activity; when the user's booking time belongs to the gathering activity time period, executing subsequent steps;

[0034] S322: Calculate the predicted number of people for the activity according to the type of the gathering activity;

[0035] S323: Calculate the room demand according to the predicted number of people for the activity;

[0036] S324: Generate an activity prediction area according to the gathering activity location; when the target hotel is within the activity prediction area, execute subsequent steps;

[0037] S325: Calculate the number of hotels in the activity prediction area;

[0038] S326: Calculate the average hotel room booking increment according to the room demand and the number of hotels in the activity forecast area; calculate the room booking growth rate of the target hotel at the user booking time through the hotel average room booking increment.

[0039] Preferably, in step S322, the acquired gathering activity big data also includes historical gathering activity number big data;

[0040] According to the type of gathering activity, calculate the historical actual number of participants and the corresponding historical number of reported participants from the big data of the historical number of participants of the gathering activity; calculate the historical attendance rate of the current type of gathering activity according to the historical actual number of participants and the corresponding historical number of reported participants; calculate the corresponding predicted number of participants by combining the historical attendance rate of the current type of gathering activity with the number of reported participants;

[0041] Among them, the predicted number of people for the event = historical attendance rate * number of people who reported for the gathering.

[0042] Preferably, in step S322, the predicted number of people for the activity is calculated based on the type of the gathering activity and the time period of the gathering activity;

[0043] Determine which day of the gathering event the user's reservation time is based on the gathering event time period:

[0044] When the user's booking time is the first day of a gathering activity, the predicted number of people for the activity = historical attendance rate * number of people reported for the gathering activity;

[0045] When the user's reservation time is the second day of the gathering activity, the predicted number of people for the activity = (1 + A1) historical attendance rate * number of people reported for the gathering activity;

[0046] When the user's reservation time is the third day of the gathering activity or later, the predicted number of people for the activity = (1 + A2) historical attendance rate * number of people reported for the gathering activity;

[0047] Among them, A1 and A2 are both constants greater than zero, and A2>A1.

[0048] Preferably, in step S321, it is determined whether the adjacent dates before and after the gathering activity time period are weekends. When the adjacent dates before and after the gathering activity time period are weekends and the user booking time belongs to the adjacent dates before and after the gathering activity time period, step S322 is continued to be executed; at this time, the logic of calculating the predicted number of people for the activity in step S322 is as follows:

[0049] Predicted number of people for the event = (1-A3) historical attendance rate * number of people reported for the gathering;

[0050] Wherein, A3 is a constant greater than zero.

[0051] Preferably, in step S321, when the gathering activity type is a concert:

[0052] Determine the main person who organizes the gathering activity; obtain the actual number of participants and the number of people who have been reported for the gathering activities in the past by the main person through the big data of the number of people in the historical activities; calculate the historical growth multiple of the number of the main person through the actual number of participants and the number of people who have been reported for the historical gathering activities; calculate the corresponding predicted number of people for the activity based on the number of people who have been reported for the gathering activities and the historical growth multiple of the number of the main person;

[0053] Predicted number of participants in an activity = historical number growth multiple * number of participants reporting for the gathering.

[0054] Preferably, in step S326, the actual distance between the target hotel and the gathering activity location is calculated; a corresponding distance factor is selected according to the actual distance, and then the room reservation growth rate of the target hotel is calculated in combination with the average room reservation increment of the hotel;

[0055] Room booking growth rate = (1 + A4) * average hotel room booking increase;

[0056] in:

[0057] When the actual distance is less than one kilometer, select A4 as R1;

[0058] When the actual distance is greater than or equal to one kilometer and less than two kilometers, select A4 as R2;

[0059] When the actual distance is greater than or equal to two kilometers and less than three kilometers, select A4 as R3;

[0060] When the actual distance is greater than or equal to three kilometers and less than four kilometers, select A4 as R4;

[0061] When the actual distance is greater than or equal to four kilometers and less than five kilometers, select A4 as R5;

[0062] Among them, R1>R2>R3>R4>R5.

[0063] Compared with the prior art, the hotel room status and price prediction method based on big data in the present invention has the following beneficial effects:

[0064] The present invention obtains relevant event big data associated with the user's booking time (such as holidays, tourist peaks, gathering activities, etc.), and then more accurately predicts the hotel's room reservation growth rate at a future time (user booking time) by analyzing the relevant event big data and combining it with historical big data, which is conducive to better calculation of vacancy rate and room price in the subsequent process.

[0065] The present invention, based on the prediction of the room reservation growth rate, further calculates the vacancy rate of the target hotel at the user's reservation time according to the current remaining number of rooms of the target hotel and the room reservation growth rate. By calculating the vacancy rate, the room supply situation in a certain period in the future can be understood, and the vacancy rate is one of the important factors affecting the room price. By calculating the vacancy rate, the present invention can more accurately and flexibly predict the room price to cope with changes in market demand.

[0066] The present invention further predicts the target room predicted price of the user's reservation time according to the current price and the vacancy rate of the target room on the basis of calculating the vacancy rate. By comprehensively considering the current price and the vacancy rate at the future time (user reservation time), a more reasonable future room price can be predicted as a reference price. On the one hand, the user can bid based on the reference price, and can use the reference price as the expected price to judge whether the merchant's counter-offer is reasonable, which not only helps to improve the success rate of hotel room bargaining, but also helps to improve the user's satisfaction, because the user is more likely to accept the price that is consistent with their expectations. On the other hand, the reference price fully considers the impact of related events (such as holidays, tourist peaks, gathering activities, etc.) at the future time (user reservation time) on the room demand, and estimates the future room demand, that is, the vacancy rate, so that the merchant can give accurate pricing and counter-offer, which helps to assist the bargaining with the user, thereby improving the accuracy and efficiency of the hotel room bargaining.

[0067] When the user's reservation time includes multiple dates, the present invention calculates the predicted room price for each date respectively, and finally generates the predicted room price change trend within the user's reservation time based on the predicted room price of each date. First, by predicting the predicted room price on a daily basis, the factors that may affect the room price on each day, such as holidays, tourist peaks, gathering activities, etc., can be captured more carefully, so as to more accurately predict the room price on each day. Secondly, the user can see the detailed change trend of the room price within the reservation time span, which helps the user to make a decision, such as the user may choose to book on a day with a lower price, or adjust the reservation time to avoid the high price period; and providing the price change trend can also increase the user's trust in the hotel's price strategy, thereby improving the user's satisfaction. Finally, by predicting the price on a daily basis and generating the change trend, the hotel's merchants can more accurately predict the future room demand situation, so as to better give pricing and counter-offer, which can not only improve the accuracy and efficiency of the hotel room bargaining, but also help the hotel reduce customer loss and revenue loss caused by shortage or surplus of rooms, and improve operational efficiency and service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0069] Figure 1 This is a logical block diagram of the hotel room status and price prediction method based on big data. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but only represents selected embodiments of the present invention. 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.

[0071] The following is a further detailed description through specific implementation methods:

[0072] Example:

[0073] This embodiment discloses a hotel room status and price prediction method based on big data.

[0074] like Figure 1As shown in the figure, the hotel room status and price prediction method based on big data includes:

[0075] S1: Obtain the user's room requirements and booking time;

[0076] In this embodiment, the user's room requirements include hotel location area, hotel star rating, room size, and room type (such as king-size room, standard room). The reservation time can be one day or multiple days.

[0077] S2: Match the target room that meets the user's room requirements and booking time; obtain the current price of the target room and the remaining number of rooms in the target hotel that includes the target room;

[0078] S3: Obtain relevant event big data associated with the user's booking time, and predict the room booking growth rate of the target hotel at the user's booking time based on the relevant event big data;

[0079] S4: Calculate the vacancy rate of the target hotel at the user's booking time according to the current remaining number of rooms of the target hotel and the predicted room booking growth rate at the user's booking time;

[0080] S5: predicting the target room predicted price at the user's booking time based on the current price of the target room and the vacancy rate of the target hotel at the user's booking time;

[0081] S6: When the user's reservation time includes multiple dates, execute steps S2 to S5 for each date to predict and obtain the target room predicted price for each date; based on the target room predicted price for each date, generate the target room predicted price change trend within the user's reservation time.

[0082] In this embodiment, when the user's reservation time includes multiple dates, the reservation time includes the check-in time, the check-out time and the number of days of stay. The predicted price change trend of the target room within the user's reservation time can be displayed by generating a line graph.

[0083] The present invention obtains relevant event big data associated with the user's booking time (such as holidays, tourist peaks, gathering activities, etc.), and then more accurately predicts the hotel's room reservation growth rate at a future time (user booking time) by analyzing the relevant event big data and combining it with historical big data, which is conducive to better calculation of vacancy rate and room price in the subsequent process.

[0084] The present invention, based on the prediction of the room reservation growth rate, further calculates the vacancy rate of the target hotel at the user's reservation time according to the current remaining number of rooms of the target hotel and the room reservation growth rate. By calculating the vacancy rate, the room supply situation in a certain period in the future can be understood, and the vacancy rate is one of the important factors affecting the room price. By calculating the vacancy rate, the present invention can more accurately and flexibly predict the room price to cope with changes in market demand.

[0085] The present invention further predicts the target room predicted price of the user's reservation time according to the current price and the vacancy rate of the target room on the basis of calculating the vacancy rate. By comprehensively considering the current price and the vacancy rate at the future time (user reservation time), a more reasonable future room price can be predicted as a reference price. On the one hand, the user can bid based on the reference price, and can use the reference price as the expected price to judge whether the merchant's counter-offer is reasonable, which not only helps to improve the success rate of hotel room bargaining, but also helps to improve the user's satisfaction, because the user is more likely to accept the price that is consistent with their expectations. On the other hand, the reference price fully considers the impact of related events (such as holidays, tourist peaks, gathering activities, etc.) at the future time (user reservation time) on the room demand, and estimates the future room demand, that is, the vacancy rate, so that the merchant can give accurate pricing and counter-offer, which helps to assist the bargaining with the user, thereby improving the accuracy and efficiency of the hotel room bargaining.

[0086] When the user's reservation time includes multiple dates, the present invention calculates the predicted room price for each date respectively, and finally generates the predicted room price change trend within the user's reservation time based on the predicted room price of each date. First, by predicting the predicted room price on a daily basis, the factors that may affect the room price on each day, such as holidays, tourist peaks, gathering activities, etc., can be captured more carefully, so as to more accurately predict the room price on each day. Secondly, the user can see the detailed change trend of the room price within the reservation time span, which helps the user to make a decision, such as the user may choose to book on a day with a lower price, or adjust the reservation time to avoid the high price period; and providing the price change trend can also increase the user's trust in the hotel's price strategy, thereby improving the user's satisfaction. Finally, by predicting the price on a daily basis and generating the change trend, the hotel's merchants can more accurately predict the future room demand situation, so as to better give pricing and counter-offer, which can not only improve the accuracy and efficiency of the hotel room bargaining, but also help the hotel reduce customer loss and revenue loss caused by shortage or surplus of rooms, and improve operational efficiency and service quality.

[0087] In order to better introduce the technical solution of the present invention, this embodiment is described through the following parts.

[0088] 1. Holiday Big Data

[0089] In actual application scenarios, people will choose to travel or visit relatives and friends during holidays. The demand for hotel rooms in many regions, especially in popular tourist cities, will increase significantly, and the corresponding room prices will also increase. Therefore, this application fully considers the dimension of holiday big data when calculating the room reservation growth rate.

[0090] In this embodiment, when the relevant event big data is holiday big data, the room reservation growth rate is predicted through the following steps:

[0091] S301: The acquired holiday big data includes holiday types and holiday time periods, as well as historical room reservation big data of all hotels during holidays;

[0092] In this embodiment, the holiday types include New Year's Day, Qingming Festival, Labor Day, Dragon Boat Festival, Mid-Autumn Festival and National Day. The holiday time period shall be based on the actual holiday time.

[0093] The historical room reservation big data here includes hotel name, holiday type, daily room reservations and room vacancy during holiday time periods. For example, xx Seven Days Hotel; New Year's Day; January 1: 65 room reservations, 35 room vacancy; January 2: 68 room reservations, 32 room vacancy; January 3: 70 room reservations, 30 room vacancy.

[0094] S302: Determine the overlapping date between the user's reservation time and the holiday time period;

[0095] In this embodiment, the overlapping date refers to a single date that is the same as the user's reservation time and the holiday time period. For example, if the user's reservation time is May 1, and the Labor Day holiday time period is from May 1 to May 5, the overlapping date is May 1. It should be noted that in this application, the user's reservation time is regarded as a single day for single calculation. When the user's reservation time involves multiple days, multiple single-day calculations need to be performed separately.

[0096] S304: Obtain historical room reservation big data of other hotels on this type of holiday, and filter out historical room reservation rates of other hotels on overlapping dates;

[0097] S304: Obtain historical room reservation big data of other hotels on this type of holiday, and then filter out historical room reservation rates of other hotels on overlapping dates;

[0098] S305: Calculate the room reservation growth rate of the target hotel at the user's reservation time based on the historical room reservation rates of the target hotel and other hotels on overlapping dates.

[0099] In this embodiment, the room reservation growth rate = 0.8 * the historical room reservation rate of the target hotel on the overlapping date + 0.2 * the historical room reservation rate of other hotels on the overlapping date. For example, if the historical room reservation rate of the target hotel on the overlapping date is 60%, and the historical room reservation rate of other hotels on the overlapping date is 80%, then the room reservation growth rate = 0.8 * 60% + 0.2 * 80% = 64%.

[0100] The present invention predicts the room reservation growth rate of the hotel at a future moment (user booking time) based on holiday big data. In actual application scenarios, affected by holiday travel, the demand for hotel rooms in many regions, especially in popular tourist cities, will increase significantly, and the corresponding room prices will also increase. Therefore, the present invention fully considers the historical room reservation rate of the hotel during the same period of holidays in previous years to predict the room reservation growth rate at future moments, which is conducive to the subsequent more accurate calculation of vacancy rate and room price. At the same time, the present invention also considers the historical room reservation rate of other hotels during the same period of holidays in previous years to assist in predicting the future room reservation growth rate, because the historical room reservation rate of other hotels can also reflect the hotel room reservation trend and demand during holidays, and then can make certain corrections to the room reservation growth rate of the target hotel, thereby improving the accuracy of the room reservation growth rate and the subsequent vacancy rate and room prediction price calculation.

[0101] 2. Big Data on Tourism Population

[0102] In actual application scenarios, there are several months of peak tourism each year, during which hotel room demand and prices will surge. Therefore, this application fully considers the dimension of holiday big data when calculating the room reservation growth rate. And unlike holidays, the peak tourism time period is only a rough interval, and the time is not fixed. Take the peak tourism period in spring as an example. Last year, the weather was better (or other factors), and people started traveling on April 15. But this year, due to bad weather (or other reasons), people did not start traveling until April 30. At this time, if you simply correspond April 15 this year with April 15 last year, the data obtained is inaccurate, and April 30 this year should be corresponded to April 15 last year.

[0103] In view of the above problem, when the relevant event big data is tourism population big data, the room reservation growth rate is predicted through the following steps:

[0104] S311: The acquired tourist population big data includes tourist peak types and tourist peak time periods, as well as historical foreign population flow big data and historical room reservation big data of all hotels during tourist peaks;

[0105] In this embodiment, the tourism peak types include four types: spring tourism peak period, summer tourism peak period, autumn tourism peak period and winter tourism peak period. The spring tourism peak period is between April 15 and May 15 each year; the summer tourism peak period is between July 15 and August 15 each year; the autumn tourism peak period is from September to October each year; and the winter tourism peak period is from December to February of the following year. Holidays such as New Year's Day, Spring Festival, Qingming Festival, Labor Day, Dragon Boat Festival, Mid-Autumn Festival, National Day and Spring Festival are not included in the tourism peak period.

[0106] It should be noted that although four peak tourist periods are set, a city may not include all four peak tourist periods, and needs to be set according to the specific situation of the city. For example, northern cities such as Harbin may only include winter peak tourist periods, while southern cities such as Chongqing and Chengdu may only include spring peak tourist periods, summer peak tourist periods, autumn peak tourist periods, etc.

[0107] The historical room reservation big data here includes hotel name, tourism peak type, daily room reservations and room vacancy during tourism peak period. For example, xx Seven Days Hotel; Spring Tourism Peak; 20xx, April 15: 30 room reservations, 70 room vacancy; April 16: 35 room reservations, 65 room vacancy; April 17: 40 room reservations, 60 room vacancy, ... and so on.

[0108] The historical foreign population flow big data here includes the peak tourism date and the corresponding foreign population flow. For example, during the peak tourism season in spring 20xx, on April 15, the foreign population flow was 500,000 people, on April 16: the foreign population flow was 700,000 people, on April 17: the foreign population flow was 1 million people, and so on.

[0109] S312: When the current date is a tourist peak period, obtain the inbound population flow of the current date;

[0110] In this embodiment, the big data of foreign population flow can be obtained through the purchase data of ticket purchasing apps such as air tickets, high-speed rail tickets, train tickets, and bus tickets, as well as the security inspection data of stations such as airports, high-speed rail stations, railway stations, and bus stations.

[0111] S313: matching the peak tourist date with the closest historical inbound population flow to the current date from the peak tourist time period according to the inbound population flow on the current date;

[0112] In this embodiment, the applicant found through analyzing the big data of tourist population that although the start time of tourism each year cannot be completely determined, the number of tourists each year changes periodically. For example, when the number of foreign population in the city reaches 500,000, it means that people have started to travel. After people start to travel, the daily flow of foreign population corresponds to the flow of foreign population in the same period last year or the previous few years (the same period here does not refer to the correspondence of dates, but to the dates corresponding to the number of foreign population, such as the previous example that April 30 of this year corresponds to April 15 of last year). Therefore, the method of determining the peak tourist time period and peak tourist date by the flow of foreign population in this embodiment is more accurate than the method of directly corresponding the date.

[0113] For example, the inflow of foreign population on the current date (April 20) is 550,000 people, while in the historical inflow of foreign population big data during the peak tourism period in spring, on April 15 last year: the inflow of foreign population was 500,000 people, on April 16: the inflow of foreign population was 700,000 people, on April 17: the inflow of foreign population was 1 million people, and so on. At this time, the peak tourism date corresponding to the current date (April 20) is April 15 last year.

[0114] S314: Calculate the date interval N between the current date and the user's reservation time; push back the travel peak date closest to the current date by N days as the target travel peak date corresponding to the user's reservation time;

[0115] In this embodiment, the user's reservation time is May 20, the current date is May 10, and the closest tourist peak date obtained by matching the current date's foreign population flow is April 20 last year. The date interval N between the current date and the user's reservation time is 10 days, so the target tourist peak date corresponding to the user's reservation time, i.e., May 20, is April 20 last year, which is pushed back 10 days, i.e., April 30 last year.

[0116] S315: Obtain the historical room reservation rate of the target hotel on the target travel peak date corresponding to the user's reservation time;

[0117] S316: Obtain the historical room reservation rate of the target travel peak date corresponding to the reservation time of other hotel sub-users;

[0118] S317: Calculate the room reservation growth rate of the target hotel at the user's reservation time based on the historical room reservation rates of the target hotel and other hotels at the user's reservation time.

[0119] In this embodiment, the room reservation growth rate = 0.8 * the historical room reservation rate of the target hotel at the time of the user's reservation + 0.2 * the historical room reservation rate of other hotels at the time of the user's reservation. For example, if the historical room reservation rate of the target hotel at the time of the user's reservation is 65%, and the historical room reservation rate of other hotels at the time of the user's reservation is 68%, then the room reservation growth rate = 0.8 * 65% + 0.2 * 68% = 65.6%.

[0120] The present invention predicts the room reservation growth rate of the target hotel at the user's booking time based on the big data of the tourist population. In actual application scenarios, the tourist peak will have a greater impact on the demand and price of hotel rooms. Therefore, the present invention fully considers the historical room reservation rate of the hotel during the same period of the tourist peak in previous years to predict the room reservation growth rate at future times, which is conducive to the subsequent more accurate calculation of vacancy rate and room price. At the same time, the present invention also considers the historical room reservation rate of other hotels during the same period of the tourist peak in previous years to assist in predicting the future room reservation growth rate, because the historical room reservation rate of other hotels can also reflect the hotel room reservation trend and demand during the tourist peak, and then can make certain corrections to the room reservation growth rate of the target hotel, thereby improving the accuracy of the room reservation growth rate and the subsequent vacancy rate and room prediction price calculation.

[0121] And unlike holidays, the peak travel time period is just a rough interval, and the time is not fixed. Take the peak travel time period in spring as an example. Last year, the weather was relatively good (or other factors), and people started traveling on April 15. But this year, due to bad weather (or other reasons), people did not start traveling until April 30. At this time, if April 15 of this year is simply corresponded with April 15 of last year, the data obtained is inaccurate, and April 30 of this year should be corresponded with April 15 of last year. In response to this problem, the present invention determines the peak travel time period and peak travel date corresponding to the user's scheduled time by the flow of foreign population, rather than directly corresponding the current date with the historical date, so that the tourist population big data can be better utilized to accurately calculate the room reservation growth rate, thereby improving the accuracy of subsequent vacancy rate and room forecast price calculation.

[0122] 3. Gathering Activity Big Data

[0123] In actual application scenarios, when a gathering event (such as sports games, music festivals, concerts, bonfire parties, etc.) is held in a city, it will have a significant impact on the demand for hotel rooms in the area in the short term. Therefore, this application fully considers the dimension of gathering event big data when calculating the room reservation growth rate.

[0124] The applicant found that the hotel room booking situation during holidays and tourist peaks is cyclical, that is, the hotel room booking situation during holidays and tourist peaks is similar every year, so we can calculate the room booking growth rate by screening the historical hotel booking big data of the same period. However, unlike holidays and tourist peaks, gathering activities do not have a fixed time and are often sudden, and the hotel room demand for gathering activities is greatly affected by the location of the gathering activities and the scale of the gathering activities (the number of participants). Users tend to book hotel rooms closer to the gathering activity location, and the more people participate in the gathering activities, the higher the demand for hotel rooms.

[0125] In view of the above problems, when the relevant event big data is aggregated activity big data, the room reservation growth rate is predicted through the following steps:

[0126] S321: The acquired gathering activity big data includes the gathering activity type, gathering activity location, gathering activity time period and the number of people who have reported for the gathering activity; when the user's booking time belongs to the gathering activity time period, executing subsequent steps;

[0127] In this embodiment, the types of gathering activities include sports games (football, basketball, tennis, table tennis, etc.), music festivals, concerts, bonfire parties, etc.

[0128] S322: Calculate the predicted number of people for the activity according to the type of the gathering activity;

[0129] S323: Calculate the room demand according to the predicted number of people for the activity;

[0130] In this embodiment, the room demand is calculated based on the allocation logic of one required room for every two people, that is, the room demand = the predicted number of people for the activity / 2.

[0131] S324: Generate an activity prediction area according to the gathering activity location; when the target hotel is within the activity prediction area, execute subsequent steps;

[0132] In this embodiment, a circular activity prediction area is generated with the gathering activity location as the center and a radius of five kilometers.

[0133] S325: Calculate the number of hotels in the activity prediction area;

[0134] S326: Calculate the average hotel room booking increment according to the room demand and the number of hotels in the activity forecast area; calculate the room booking growth rate of the target hotel at the user booking time through the hotel average room booking increment.

[0135] In this embodiment, the average hotel room booking increment = room demand / number of hotels.

[0136] The present invention predicts the room reservation growth rate of the target hotel at the user's booking time based on the big data of the gathering activities. In the actual application scenario, when a gathering activity is held in a city, it will have a significant impact on the hotel room demand in the city in the short term. Therefore, the present invention takes into account the average hotel reservation increment of the hotel under the gathering activity, and then accurately predicts the room reservation growth rate at future times, which is conducive to the subsequent more accurate calculation of vacancy rate and room price. At the same time, the present invention first generates an activity prediction area according to the gathering activity location, then calculates the predicted number of activities and the room demand, and finally calculates the number of hotels in the activity prediction area and calculates the average hotel reservation increment in combination with the room demand, so that the gathering activity location and the predicted number of activities (number of participants) can be used as the main considerations to calculate the room reservation growth rate, so that the room reservation growth rate considering the big data of the gathering activities can be accurately calculated.

[0137] 1. Consider the attendance rate

[0138] In actual application scenarios, cluster activities will have a number of people who have reported for the cluster activities (such as the number of people that a venue for a sports game or concert can accommodate). However, in many cases, cluster activities are not always full. Some cluster activities even have a relatively low attendance rate. In this case, if subsequent calculations are made directly based on the number of people who have reported for the cluster activities, there will be large errors.

[0139] In view of the above problems, the gathering activity big data obtained in this embodiment also includes the historical activity number big data of the gathering activity; according to the gathering activity type, the historical actual number of participants and the corresponding historical gathering activity reported number of participants are calculated from the historical activity number big data of the gathering activity; the historical attendance rate of the current gathering activity type is calculated according to the historical actual number of participants and the corresponding historical gathering activity reported number of participants; the corresponding activity predicted number of participants is calculated by combining the historical attendance rate of the current gathering activity type with the number of gathering activity reported number of participants;

[0140] Among them, the predicted number of people for the event = historical attendance rate * number of people who reported for the gathering.

[0141] The present invention calculates the historical attendance rate of this type of gathering activity based on the big data of historical attendance of gathering activities, and then calculates the corresponding predicted attendance of the activity through the historical attendance rate of this type of gathering activity combined with the number of people who have reported for the gathering activity. This can effectively consider the attendance situation of different types of gathering activities, that is, the actual number of participants, so that the subsequent room reservation growth rate can be calculated based on a more accurate predicted number of activities.

[0142] 2. Consider the user continues to play

[0143] In actual application scenarios, people often do not leave the city immediately after participating in a gathering activity, and a large number of them will continue to play in the city. In other words, when a gathering interaction needs to last for several days, the actual number of people on the second day of the gathering activity may be higher than the predicted number we calculated based on the gathering activity type, because the second day of the gathering activity also includes users who did not leave on the first day, and so on, the third day of the gathering activity will also include users who did not leave on the first and second days. At this time, it is inaccurate to directly use the predicted number of people calculated based on the gathering activity type as the predicted number of people for the activity.

[0144] In view of the above problems, in this embodiment, the predicted number of people for an activity is calculated based on the type of gathering activity and the time period of the gathering activity;

[0145] Determine which day of the gathering event the user's reservation time is based on the gathering event time period:

[0146] When the user's booking time is the first day of a gathering activity, the predicted number of people for the activity = historical attendance rate * number of people reported for the gathering activity;

[0147] When the user's reservation time is the second day of the gathering activity, the predicted number of people for the activity = (1 + A1) historical attendance rate * number of people reported for the gathering activity;

[0148] When the user's reservation time is the third day of the gathering activity or later, the predicted number of people for the activity = (1 + A2) historical attendance rate * number of people reported for the gathering activity;

[0149] Among them, A1 and A2 are both constants greater than zero, and A2>A1.

[0150] When calculating the predicted number of people for an activity, the present invention not only takes into account the attendance rate corresponding to the type of gathering activity, but also further takes into account the situation in which users do not leave after participating in the gathering activity and continue to play. At this time, the demand for hotel rooms is still very high, so the predicted number of people for an activity can be calculated more accurately through reasonable design of A1 and A2, and then the subsequent room reservation growth rate can be calculated based on the more accurate predicted number of people for the activity.

[0151] 3. Consider the weekend

[0152] In actual application scenarios, the applicant found that when the adjacent dates before and after the gathering activity time period are weekends, some users will come to the city to play in advance or continue to play after the gathering activity ends, that is, spend the weekend in the city. In this case, even if the user's booking time does not belong to the gathering activity time period, as long as it belongs to the weekend before and after the gathering activity time period, it will be affected to a certain extent by the gathering activity.

[0153] Therefore, in this embodiment, it is further determined whether the adjacent dates before and after the gathering activity time period are weekends. When the adjacent dates before and after the gathering activity time period are weekends and the user's reservation time belongs to the adjacent dates before and after the gathering activity time period, step S322 is continued to be executed; at this time, the logic of calculating the predicted number of people for the activity in step S322 is as follows:

[0154] Predicted number of people for the event = (1-A3) historical attendance rate * number of people reported for the gathering;

[0155] Wherein, A3 is a constant greater than zero.

[0156] When the adjacent dates before and after the gathering activity time period are weekends and the user's booking time belongs to the adjacent dates before and after the gathering activity time period, the present invention will also perform the calculation of the room reservation growth rate. This is to take into account the situation where users spend weekends in the city due to the gathering activities, so that the impact of the dates before and after the gathering activities on the hotel room prices can be fully considered, so that the subsequent room reservation growth rate can be calculated based on a more accurate predicted number of activities.

[0157] 4. Consider fan support for concerts

[0158] In actual application scenarios, when some well-known stars (key personnel) hold concerts, the fans of the key personnel will provide fan support. At this time, the number of people participating in the support may even be far greater than the number of people who have reported for the gathering. If subsequent calculations are made directly based on the number of people who have reported for the gathering, there will be large errors.

[0159] To address this issue, in this embodiment, when the gathering activity type is a concert:

[0160] Determine the main person who organizes the gathering activity; obtain the actual number of participants and the number of people who have been reported for the gathering activities in the past by the main person through the big data of the number of people in the historical activities; calculate the historical growth multiple of the number of the main person through the actual number of participants and the number of people who have been reported for the historical gathering activities; calculate the corresponding predicted number of people for the activity based on the number of people who have been reported for the gathering activities and the historical growth multiple of the number of the main person;

[0161] Predicted number of participants in an activity = historical number growth multiple * number of participants reporting for the gathering.

[0162] When facing a gathering activity such as a concert, the present invention first determines the main personnel who organize the gathering activity, and then obtains the historical actual number of participants and the historical number of reported gathering activities when the main personnel organized the gathering activities in the past through the big data of historical activity numbers, and then calculates the historical number growth multiple of the main personnel (that is, the number of fans participating in the support) through the historical actual number of participants and the historical number of reported gathering activities, and finally calculates the corresponding activity predicted number of people according to the number of reported gathering activities and the historical number growth multiple of the main personnel, which can effectively take into account the situation of fan support for gathering activities such as concerts, so that the subsequent room reservation growth rate can be calculated based on a more accurate activity predicted number of people.

[0163] 5. Consider distance

[0164] In actual application scenarios, when participating in gathering activities, for the convenience of travel, users often give priority to booking rooms in hotels closest to the gathering activity venue. Therefore, it is necessary to consider the actual distance between the hotel and the gathering activity venue when calculating the room reservation growth rate.

[0165] In this embodiment, the actual distance between the target hotel and the gathering activity location is calculated; the corresponding distance factor is selected according to the actual distance, and then the room reservation growth rate of the target hotel is calculated in combination with the average room reservation increment of the hotel;

[0166] Room booking growth rate = (1 + A4) * average hotel room booking increase;

[0167] in:

[0168] When the actual distance is less than one kilometer, select A4 as R1;

[0169] When the actual distance is greater than or equal to one kilometer and less than two kilometers, select A4 as R2;

[0170] When the actual distance is greater than or equal to two kilometers and less than three kilometers, select A4 as R3;

[0171] When the actual distance is greater than or equal to three kilometers and less than four kilometers, select A4 as R4;

[0172] When the actual distance is greater than or equal to four kilometers and less than five kilometers, select A4 as R5;

[0173] Among them, R1>R2>R3>R4>R5. R4 and R5 can be set to negative numbers as needed.

[0174] The present invention first calculates the actual distance between the target hotel and the gathering activity location of the gathering activity, and then selects the corresponding preset distance factor through the actual gathering to calculate the room reservation growth rate of the target hotel. In this way, the influence of the actual distance between the hotel and the gathering activity location on the user's choice of hotel can be effectively considered. In this way, the big data of gathering activities can be better utilized to accurately calculate the room reservation growth rate, thereby improving the accuracy of subsequent vacancy rate and room forecast price calculation.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. A hotel room status and price prediction method based on big data, characterized in that: include: S1: Obtain the user's room requirements and booking time; S2: Match the target room that meets the user's room requirements and booking time; obtain the current price of the target room and the remaining number of rooms in the target hotel that includes the target room; S3: Obtain relevant event big data associated with the user's booking time, and predict the room booking growth rate of the target hotel at the user's booking time based on the relevant event big data; S4: Calculate the vacancy rate of the target hotel at the user's booking time according to the current remaining number of rooms of the target hotel and the predicted room booking growth rate at the user's booking time; S5: Predict the target room price at the user's booking time based on the current price of the target room and the vacancy rate of the target hotel at the user's booking time.

2. The hotel room status and price prediction method based on big data as claimed in claim 1, characterized in that: Also includes step S6; S6: When the user's reservation time includes multiple dates, execute steps S2 to S5 for each date to predict and obtain the target room predicted price for each date; based on the target room predicted price for each date, generate the target room predicted price change trend within the user's reservation time.

3. The hotel room status and price prediction method based on big data as claimed in claim 1, characterized in that: In step S3, the relevant event big data includes holiday big data; Forecast room booking growth rate by following these steps: S301: The acquired holiday big data includes holiday types and holiday time periods, as well as historical room reservation big data of all hotels during holidays; S302: Determine the overlapping date between the user's reservation time and the holiday time period; S303: Obtain historical room reservation big data of the target hotel on the holiday type, and filter out the historical room reservation rate of the target hotel on the overlapping dates; S304: Obtain historical room reservation big data of other hotels on this type of holiday, and filter out historical room reservation rates of other hotels on overlapping dates; S305: Calculate the room reservation growth rate of the target hotel at the user's reservation time based on the historical room reservation rates of the target hotel and other hotels on overlapping dates.

4. The hotel room status and price prediction method based on big data as claimed in claim 1, characterized in that: In step S3, the relevant event big data includes tourist population big data; Forecast room booking growth rate by following these steps: S311: The acquired tourist population big data includes tourist peak types and tourist peak time periods, as well as historical foreign population flow big data and historical room reservation big data of all hotels during tourist peaks; S312: When the current date is a tourist peak period, obtain the inbound population flow of the current date; S313: matching the historical foreign population flow rate with the foreign population flow rate on the current date from the tourist peak time period according to the historical foreign population flow rate big data; S314: Calculate the interval N between the current date and the user's reservation time; push back N days the tourist peak date with the closest inflow of foreign people to the current date as the target tourist peak date corresponding to the user's reservation time; S315: Obtain the historical room reservation rate of the target hotel on the target travel peak date corresponding to the user's reservation time; S316: Obtain the historical room reservation rate of the target travel peak date corresponding to the reservation time of other hotel sub-users; S317: Calculate the room reservation growth rate of the target hotel at the user's reservation time based on the historical room reservation rates of the target hotel and other hotels at the user's reservation time.

5. The hotel room status and price prediction method based on big data as claimed in claim 1, characterized in that: In step S3, the relevant event big data includes the aggregation activity big data; Forecast room booking growth rate by following these steps: S321: The acquired gathering activity big data includes the gathering activity type, gathering activity location, gathering activity time period and the number of people who have reported for the gathering activity; when the user's booking time belongs to the gathering activity time period, executing subsequent steps; S322: Calculate the predicted number of people for the activity according to the type of the gathering activity; S323: Calculate the room demand according to the predicted number of people for the activity; S324: Generate an activity prediction area according to the gathering activity location; when the target hotel is within the activity prediction area, execute subsequent steps; S325: Calculate the number of hotels in the activity prediction area; S326: Calculate the average hotel room booking increment according to the room demand and the number of hotels in the activity forecast area; calculate the room booking growth rate of the target hotel at the user booking time through the hotel average room booking increment.

6. The hotel room status and price prediction method based on big data as claimed in claim 5, characterized in that: In step S322, the acquired gathering activity big data also includes historical gathering activity number big data; According to the type of gathering activity, calculate the actual historical number of participants in the current type of gathering activity and the corresponding number of historical gathering activity reports from the big data of the historical number of participants in the gathering activity; Calculate the historical attendance rate of the current gathering activity type based on the actual number of historical participants and the corresponding number of historical gathering activity reports; calculate the corresponding activity forecast number based on the historical attendance rate of the current gathering activity type combined with the number of gathering activity reports; Among them, the predicted number of people for the event = historical attendance rate * number of people who reported for the gathering.

7. The hotel room status and price prediction method based on big data as claimed in claim 6, characterized in that: In step S322, the predicted number of people for the activity is calculated based on the type of the gathering activity and the time period of the gathering activity; Determine which day of the gathering event the user's reservation time is based on the gathering event time period: When the user's booking time is the first day of a gathering activity, the predicted number of people for the activity = historical attendance rate * number of people reported for the gathering activity; When the user's reservation time is the second day of the gathering activity, the predicted number of people for the activity = (1 + A1) historical attendance rate * number of people reported for the gathering activity; When the user's reservation time is the third day of the gathering activity or later, the predicted number of people for the activity = (1 + A2) historical attendance rate * number of people reported for the gathering activity; Among them, A1 and A2 are both constants greater than zero, and A2>A1.

8. The hotel room status and price prediction method based on big data as claimed in claim 6, characterized in that: In step S321, it is determined whether the adjacent dates before and after the gathering activity time period are weekends. When the adjacent dates before and after the gathering activity time period are weekends and the user's booking time belongs to the adjacent dates before and after the gathering activity time period, step S322 is continued to be executed; at this time, the logic of calculating the predicted number of people for the activity in step S322 is as follows: Predicted number of people for the event = (1-A3) historical attendance rate * number of people reported for the gathering; Wherein, A3 is a constant greater than zero.

9. The hotel room status and price prediction method based on big data as claimed in claim 5, characterized in that: In step S321, when the gathering activity type is a concert: Determine the main person who organizes the gathering activity; obtain the actual number of participants and the number of people who have been reported for the gathering activities in the past by the main person through the big data of the number of people in the historical activities; calculate the historical growth multiple of the number of the main person through the actual number of participants and the number of people who have been reported for the historical gathering activities; calculate the corresponding predicted number of people for the activity based on the number of people who have been reported for the gathering activities and the historical growth multiple of the number of the main person; Predicted number of participants in an activity = historical number growth multiple * number of participants reporting for the gathering.

10. The hotel room status and price prediction method based on big data as claimed in claim 5, characterized in that: In step S326, the actual distance between the target hotel and the gathering activity location is calculated; the corresponding distance factor is selected according to the actual distance, and then the room reservation growth rate of the target hotel is calculated in combination with the average room reservation increment of the hotel; Room booking growth rate = (1 + A4) * average hotel room booking increase; in: When the actual distance is less than one kilometer, select A4 as R1; When the actual distance is greater than or equal to one kilometer and less than two kilometers, select A4 as R2; When the actual distance is greater than or equal to two kilometers and less than three kilometers, select A4 as R3; When the actual distance is greater than or equal to three kilometers and less than four kilometers, select A4 as R4; When the actual distance is greater than or equal to four kilometers and less than five kilometers, select A4 as R5; Among them, R1>R2>R3>R4>R5.

Images