CN106780173B - OTA hotel inventory management method and system - Google Patents
OTA hotel inventory management method and system Download PDFInfo
- Publication number
- CN106780173B CN106780173B CN201611093188.4A CN201611093188A CN106780173B CN 106780173 B CN106780173 B CN 106780173B CN 201611093188 A CN201611093188 A CN 201611093188A CN 106780173 B CN106780173 B CN 106780173B
- Authority
- CN
- China
- Prior art keywords
- house
- hotel
- inventory
- ota
- closing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000007726 management method Methods 0.000 title claims abstract description 33
- 238000012549 training Methods 0.000 claims description 23
- 238000000034 method Methods 0.000 claims description 9
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000007477 logistic regression Methods 0.000 claims description 7
- 239000013589 supplement Substances 0.000 abstract description 3
- 230000007812 deficiency Effects 0.000 abstract 1
- 239000012452 mother liquor Substances 0.000 description 11
- 238000003066 decision tree Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000008774 maternal effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004308 accommodation Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/12—Hotels or restaurants
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/02—Reservations, e.g. for tickets, services or events
Landscapes
- Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an OTA hotel inventory management method and system, wherein the management method comprises the following steps: predicting the closing time length of each house type of the hotel on the OTA on different living days and whether a demand gap exists during closing; for the house type with the demand gap, predicting whether a hotel party has residual stock; issuing an offer to the hotel party requesting an increase in the OTA of the house-type inventory for which there is a demand gap, upon predicting that there is inventory remaining at the hotel party. The invention makes up the deficiency of adding the inventory through manual experience in the prior art, can issue an invitation for adding the inventory to a hotel party in a targeted and automatic manner, and timely supplements the inventory of house types with demand gaps, thereby ensuring that the hotel house types have sufficient inventory for users to reserve and avoiding the excess inventory due to insufficient reservation quantity of some house types on the other hand, and having the advantage of improving the inventory management efficiency and quality of the hotel.
Description
Technical Field
The invention relates to a hotel inventory management method and system for an OTA (on-line travel agency).
Background
Booking hotel accommodations using OTA has been a common choice for people to travel or work, and the OTA is also increasingly demanding on its functionality. Due to limited hotel resources on the OTA or large bookings for hot hotels, hotel resources often report inventory depletion many days before the check-in day, resulting in the inability of users to book these hotels. In order to meet the booking requirements of users and avoid the situation that a hotel cannot be booked due to stock exhaustion in the prior art, a common solution is that service personnel of the OTA determines stock required to be increased according to individual experience, and negotiates with a hotel party to obtain stock resources.
However, there are many problems with the way to add inventory by human experience: 1. the coverage is narrow, as many as 50 million hotels exist in the whole domestic market, millions of hotels are released, and the coverage is very limited depending on business personnel; 2. future stock is added according to historical data, the target hit rate is low, and meanwhile, the quality is uneven due to the fact that the stock completely depends on personal capacity and experience of the service; 3. the fluctuation of the demand gap change caused by seasonal change cannot be responded to manually in time; 4. only for the house type which is closed, no house is related at present, but the stock which is closed in the future has no way to operate.
Disclosure of Invention
The invention aims to overcome the defects of the prior art that the method for adding the stock through manual experience has the problems, and provides an OTA hotel stock management method and system capable of accurately increasing the stock without depending on the manual experience.
The invention solves the technical problems through the following technical scheme:
an OTA hotel inventory management method is characterized by comprising the following steps:
predicting the closing time length of each house type of the hotel on the OTA on different living days and whether a demand gap exists during closing;
for the house type with the demand gap, predicting whether a hotel party has residual stock;
issuing an offer to the hotel party requesting an increase in the OTA of the house-type inventory for which there is a demand gap, upon predicting that there is inventory remaining at the hotel party.
The house closing means that new reservation cannot be added due to the fact that the reservation quantity of the house type of the hotel reaches the stock of the house type; the closing time length of one house type different from the entering date refers to the time length of closing the house before the entering date of the house type is reached. By predicting the closing time of each room type of the hotel on the OTA on different living days, service personnel can master the closing condition of each room type, and the stock of each room type can be preliminarily adjusted. By predicting whether a demand gap exists when the hotel room type is closed and whether the room type with the demand gap has residual inventory on the hotel side, an invitation for increasing the inventory is issued to the hotel side in a targeted manner, timely and automatically, on one hand, sufficient inventory of the hotel room type is ensured to be reserved by a user, and on the other hand, the situation that the inventory of some room types is excessive due to insufficient reservation quantity is avoided.
Preferably, the OTA hotel inventory management method predicts the closing time length of each house type of the hotel on the OTA on different stay days through the following steps:
the method comprises the steps of establishing a house closing time length prediction model of different entering days respectively aiming at each house type, collecting historical data of the house type as a training sample by adopting an xgboost algorithm, and selecting one or more indexes from indexes used for reflecting one or more conditions of booking conditions, browsed conditions, price conditions, tension conditions or evaluation conditions of the house type as input variables of the house closing time length prediction model.
Preferably, the OTA hotel inventory management method predicts whether a demand gap exists when each room type of the hotel on the OTA is different to enter a daily customs room by the following steps:
aiming at each house type, respectively establishing a prediction model of a warehouse closing time gap on different entering days, wherein the prediction model of the house closing time adopts an xgboost algorithm, collects historical data of the house type as a training sample, and selects one or more indexes from indexes used for reflecting one or more conditions of booking condition, browsed condition, price condition or tension condition of the house type as input variables of the prediction model of the warehouse closing time gap.
Preferably, the OTA hotel inventory management method predicts whether the inventory of the restaurant side is remained for the type of the room with the demand gap by the following steps:
and calculating the probability of the remaining stock of the house type with the demand gap by using a logistic regression algorithm, and collecting historical data of the house type as a training sample.
Preferably, the OTA hotel inventory management method further comprises:
selecting the house type with the historical night amount larger than the night amount threshold value and related house history from all house types of all hotels on the OTA;
and predicting whether the historical night time is larger than a night time threshold value, and the closing time length of the house type related to the house history on different entering days and whether a demand gap exists during closing.
An OTA hotel inventory management system is characterized by comprising:
the first prediction unit is used for predicting the closing time length of each house type of the hotel on the OTA on different living days and whether a demand gap exists during closing;
the second prediction unit is used for predicting whether the hotel party has residual stock for the house type with the demand gap;
and the issuing unit is used for issuing an offer to the hotel party when the fact that the inventory of the restaurant party is remained is predicted, wherein the offer is used for requesting to add the stock of the house type with the demand gap on the OTA.
Preferably, the first prediction unit includes:
the house closing time length prediction module is used for respectively establishing house closing time length prediction models of different entering days for each house type, the house closing time length prediction models adopt an xgboost algorithm, historical data of the house types are collected to be used as training samples, and one or more indexes are selected from indexes used for reflecting one or more conditions of booking conditions, browsed conditions, price conditions, tension conditions or evaluation conditions of the house types to be used as input variables of the house closing time length prediction models.
Preferably, the first prediction unit includes:
the system comprises a house closing inventory gap prediction module, a house closing inventory gap prediction module and a house closing inventory gap prediction module, wherein the house closing inventory gap prediction module is used for respectively establishing house closing inventory gap prediction models of different entering days for each house type, the house closing duration prediction model adopts an xgboost algorithm, historical data of the house type are collected to be used as training samples, and one or more indexes are selected from indexes used for reflecting one or more conditions of booking conditions, browsed conditions, price conditions or tension conditions of the house type to be used as input variables of the house closing inventory gap prediction model.
Preferably, the second prediction unit is configured to calculate a probability that the restaurant side has remaining stock for the house type with the demand gap by using a logistic regression algorithm, and collect historical data of the house type as a training sample.
Preferably, the OTA hotel inventory management system further comprises:
the selecting unit is used for selecting the house type with the historical night amount larger than the night amount threshold value and related house history from all house types of all hotels on the OTA;
the first prediction unit is further used for predicting the historical night time amount which is larger than the night time amount threshold value and whether a demand gap exists in the closing process or not when the house type related to the house history is different from the entering date.
On the basis of the common knowledge in the field, the above preferred conditions can be combined randomly to obtain the preferred embodiments of the invention.
The positive progress effects of the invention are as follows: the invention can issue an invitation for increasing the stock to a hotel party in a targeted and automatic manner, and timely supplement the stock of the house type with a demand gap, thereby ensuring that the house type of the hotel has sufficient stock for a user to reserve, avoiding the excess stock due to insufficient reservation quantity of some house types and having the advantage of improving the stock management efficiency and quality of the hotel.
Drawings
Fig. 1 is a flowchart of an OTA hotel inventory management method according to a preferred embodiment 1 of the present invention.
Fig. 2 is a system block diagram of the OTA hotel inventory management system according to the preferred embodiment 1 of the present invention.
Detailed Description
The invention is further illustrated by the following examples, which are not intended to limit the scope of the invention.
Example 1
An OTA hotel inventory management method, as shown in fig. 1, comprises the following steps:
and step 101, selecting the house type with the historical night amount larger than the night amount threshold value and related to the house history from all house types of all hotels on the OTA. Wherein, the larger the historical night amount of the house type is, the better the booking condition of the house type is, and the threshold value of the night amount can be freely set. Since the total house types of all hotels on the OTA may be large in number, the historical night amount is greater than the night amount threshold, and the house types of the relevant house history are better in booking situation and have a case of closing the house before, so that the possibility of closing the house again is high, the house types are used as the target house types predicted in the subsequent step in the embodiment, the workload can be reduced, and the management efficiency can be improved. If the total house types of all the hotels on the OTA are small in number or other requirements exist, the manager can select the total house types of all the hotels as the target house types predicted by the subsequent step or freely select a part of specific house types as the target house types predicted by the subsequent step. For convenience of illustration, the target house type mentioned in the subsequent step is the house type of the history of the house selected in step 101, which is greater than the threshold value of the amount of the night.
And step 102, predicting the closing time of different entrance days of the target house type and whether a demand gap exists during closing. The OTA hotel inventory management method predicts the closing time of each room type of the hotel on the OTA on different stay days through the following steps:
respectively establishing a time length of closing prediction model of different entering days aiming at each house type, namely respectively establishing a time length of closing prediction model of each entering day of each house type; the house-closing duration prediction model adopts an xgboost algorithm, collects historical data of the house type as a training sample, and selects one or more indexes from indexes used for reflecting one or more conditions of booking condition, browsed condition, price condition, tension condition or evaluation condition of the house type as input variables of the house-closing duration prediction model. Specifically, the xgboost algorithm is a decision tree algorithm based on iterative accumulation, and mainly accumulates the results of a plurality of decision trees as a final prediction result by constructing a group of weak learners. The target algorithm for Xgboost is:
wherein,
representing the predicted value of a sample i, k representing the total number of decision trees, and n representing the number of samples; y denotes the true value of the sample, Ω (f)k) Is the complexity at the kth decision. T represents the number of leaf nodes of the Kth number, omegajThe representation prediction is worth regularizing.
When training samples are collected, historical data of 202 days in history or historical data of other days can be collected as training samples.
On the selection of the input variables of the house closing time length prediction model, indexes such as a preset amount, a preset number of days, a recent output and the like capable of reflecting the booking condition of each house type are selected in a hotel order form through data analysis and visual analysis; in the browsing exterior and interior, indexes such as browsing times, browsing number, conversion rate and the like capable of reflecting the browsed conditions of all house types are selected; in the price list, house type prices capable of reflecting the price conditions of all house types, the lowest price and the highest price of a mother liquor shop and the like are selected; in the form of the chamber, the indexes such as the current chamber state, the current tension, the historical tension and the like which can reflect the tension of each chamber type are selected; in the hotel attribute list, indexes such as hotel point scores, number of people to be assessed, hotel recommendation levels and the like which can reflect the evaluation condition of each house type are selected.
After the processing of the characteristic engineering, 20 indexes are finally selected as input variables of the model. The method comprises the following steps: average house type tonus 3 days in advance to 10 days in advance; the ratio of the house type price to the median of the maternal wine shop price; the ratio of the house type price to the minimum value of the mother liquor shop price; the ratio of the average browsing times from 10 days to 20 days in advance to the average browsing times from 3 days to 10 days in advance; mean value of house type night hours 3 days in advance to 10 days in advance; a ratio of the house type advanced predetermined amount to the maternal wine shop advanced predetermined amount; a ratio of the house type advanced predetermined amount to the parent base house type advanced predetermined amount; dividing the ratio of the browsing times 3 days in advance to 10 days in advance to the preset amount in advance by the ratio of the browsing times 10 days in advance to 20 days in advance to the preset amount in advance; the ratio of the average tension of the house type predicted days in advance to the average tension of the mother liquor shop; the preset amount of the city in the online channel is 10 days to 60 days in advance; the preset amount of the city in the app channel is 10 days to 60 days in advance; average tension of 3-10 days in advance of the mother liquor shop, 11 days in advance of the city, and 12 days in advance of the city; the tension of the city 13 days ahead; the ratio of the browsing times of the mother liquor shop 10 days to 60 days in advance to the browsing times of the city 10 days to 60 days in advance; the ratio of the mean tension of 10 days to 60 days in advance of the city to the tension of 10 days to 60 days in advance of the mother liquor shop to the tension of 10 days to 60 days in advance of the house type; the house type is 4 days earlier in tension, and the house type is 10 days earlier in tension.
In the concrete model processing, on one hand, different models are established by adopting different input variables and different model parameters according to different incoming days. In the selection of the variable, the closer the date of stay is to the current date, the more dates the variable selection is input. Such as: aiming at the prediction of the closing time length of the entering-living day on the same day, the input variable adopted by the invention selects the tension degree of the entering-living day in advance of 1 day and the browsing amount of the entering-living day in advance of 1 day; aiming at the prediction of the closing time length of the stay day on the next day, the input variables adopted by the method are the tension degree of the stay day 2 days ahead, the browsing amount of the stay day 2 days ahead and the like; aiming at the prediction of the closing time length of the stay day on the third day, the input variables adopted by the method are the tension degree of the stay day 3 days ahead, the browsing amount of the stay day 3 days ahead and the like. Thus, the accuracy of the prediction model can be improved; the parameters of the day of the residence are as follows: nrounds is 12, subsample is 0.8, eta is 0.15, max _ depth is 3, objective is reg, logistic; the parameters of the next day of the stay are: nrounds is 30, subsample is 0.4, eta is 0.1, max _ depth is 4, objective is reg, logistic; the parameters of the third day of stay are: nrounds is 15, subsample is 0.3, eta is 0.15, max _ depth is 3, objective is reg, logistic; on the other hand, the invention adopts each house type as a model, thereby further improving the prediction accuracy. The influence factors of the closing time of each house type are different, such as: the closing time of some house types is greatly influenced by price, the closing time of some house types is greatly influenced by CR, and the variable importance of each house type is different, so that a single model and a single model parameter cannot meet the requirement; in addition, the model adopts a parallel mode of 3 threads to carry out operation, and the operation speed is greatly improved.
The OTA hotel inventory management method predicts whether a demand gap exists when each room type of a hotel on the OTA is different to enter a closing room on a residence day through the following steps:
aiming at each house type, respectively establishing a prediction model of a warehouse closing time gap on different entering days, wherein the prediction model of the house closing time adopts an xgboost algorithm, collects historical data of the house type as a training sample, and selects one or more indexes from indexes used for reflecting one or more conditions of booking condition, browsed condition, price condition or tension condition of the house type as input variables of the prediction model of the warehouse closing time gap. Specifically, when the training samples are collected, historical data of 202 days in history or historical data of other days can be collected as the training samples.
And finally selecting the following indexes as input variables of the model through the processing of characteristic engineering on the selection of the input variables of the inventory gap prediction model during the house closing. The method comprises the following steps: the predetermined night amount of each house type which can reflect the reservation condition of each house type is 3 days ahead, the predetermined night amount of the house type is 4 days ahead, the predetermined night amount of the mother liquor shop is 3-10 days ahead, the proportion of the night amount of the mother liquor shop 3-60 days ahead in the city advance night amount, the maximum sales volume of the house type in the past 3 months and other indexes; indexes such as the ratio of the house type price to the median of the price of the mother-foundation hotel, the ratio of the house type price to the minimum value of the price of the mother-foundation hotel and the like which can reflect the price condition of each house type; the current house state of the house type capable of reflecting the tension condition of each house type, the tension of the house type in advance of day 3-10, the average tension of the city in advance of day 3-60, the average tension of the mother liquor shop in advance of day 3-60, the ratio of the average tension of the mother liquor shop in advance of day 3-60 to the average tension of the house type in advance of day 3-60, the number of rooms of the mother hotel and other indexes; the method can reflect indexes such as the ratio of the browsing times of the liquor shop of which each house type is browsed 3-60 days in advance to a predetermined night amount of the liquor shop of which the time is 3-60 days in advance, the ratio of the browsing times of the liquor shop of which the time is 3-10 days in advance to the browsing times of the liquor shop of which the time is 10-60 days in advance, the number of people browsing the liquor shop in advance and the like.
On the one hand, different models are established by adopting different input variables and different model parameters aiming at different dates of living and different house types, so that the precision of the models is ensured, and the parameters of the models aiming at the days of living are as follows: nrounds is 15, subsample is 0.8, eta is 0.15, max _ depth is 3, objective is reg, logistic; the parameters of the next day of the stay are: nrounds is 24, subsample is 0.4, eta is 0.1, max _ depth is 4, objective is reg, logistic; the parameters of the third day of stay are: nrounds is 15, subsample is 0.3, eta is 0.15, max _ depth is 3, objective is reg, logistic; on the other hand, the algorithm adopts a mode of parallel operation of 3 threads, and the budget speed is greatly improved.
And step 103, predicting whether the storefront has residual stock for the house type with the demand gap. The OTA hotel inventory management method predicts whether the residual inventory exists in the restaurant for the type of the room with the demand gap through the following steps:
and calculating the probability of the remaining stock of the house type with the demand gap by using a logistic regression algorithm, and collecting historical data of the house type as a training sample. Specifically, the problem classified 0/1 is primarily handled by a sigmoid function, where the formula for sigmoid is:
where h (x) is a value of- ∞, + ∞, which can be transformed into a number between [0, 1] by a sigmoid function;
wirepresents the variable xiThe coefficient of (a).
On the mark of the independent variable, a certain entry date in history is successfully supplemented to the stock house type, the independent variable is set as '1', the stock house type is not successfully supplemented in history, the independent variable is set as '0', and on the selection of the dependent variable, the indexes of a preset amount in advance, the number of times of browsing in advance, the tension of the recent house type, the ratio of the price of the house type to the lowest price of the mother basic house type, the ratio of the price of the house type to the median of the mother liquor shop and the like are finally selected through characteristic engineering. Through a logistic regression algorithm, historical data of past 200 days or other days are selected as training samples, the possibility that the house type successfully supplements the inventory in the future is predicted, and the probability value is larger, so that the probability that the hotel has a house is larger. The smaller the probability, the smaller the probability of the hotel having a room, and the greater the difficulty of the competition.
And 104, when the fact that the remaining stock exists in the restaurant is predicted, issuing an offer to the hotel side, wherein the offer is used for requesting to add the stock of the house type with the demand gap on the OTA. Specifically, the offer can be issued to the hotel party when the probability of the restaurant party having a room is predicted to be greater than a preset threshold, if the offer is successful, the supplemented inventory is placed into an inventory pool of the OTA for selling so as to be purchased by the user, meanwhile, the selling condition is monitored, and if the offer is unsuccessful, manual intervention can be performed, negotiation with the hotel party is performed, and whether the inventory can be supplemented or not is determined.
An OTA hotel inventory management system of this embodiment, as shown in fig. 2, includes: the device comprises a selecting unit 201, a first prediction unit 202, a second prediction unit 203 and an issuing unit 204.
The selecting unit 201 is configured to select a room type with a historical night amount larger than a night amount threshold and related to the room history from all room types of all hotels on the OTA. For convenience of explanation, the selected historical night time amount is greater than the night time amount threshold value, and the house type related to the house history is recorded as the target house type.
The first prediction unit 202 is configured to predict the closing time of the target house type on different entrance days and whether a demand gap exists during closing. The first prediction unit 202 specifically includes:
the house closing time length prediction module 2021 is configured to establish a house closing time length prediction model for each house type, respectively, where the house closing time length prediction model uses an xgboost algorithm, collects historical data of the house type as a training sample, and selects one or more indexes from indexes used for reflecting one or more conditions of a booking condition, a browsed condition, a price condition, a tension condition, or an evaluation condition of the house type as input variables of the house closing time length prediction model.
The system comprises a closing inventory gap prediction module 2022, configured to establish a closing inventory gap prediction model for each house type on different entering days, respectively, where the closing time prediction model uses an xgboost algorithm, collects historical data of the house type as a training sample, and selects one or more indexes from indexes used for reflecting one or more conditions of a booking situation, a browsed situation, a price situation, or a tension situation of the house type as input variables of the closing inventory gap prediction model.
The second prediction unit 203 is configured to predict whether there is a remaining stock in the store for the house type with the demand gap. The second prediction unit 203 is specifically configured to calculate, by using a logistic regression algorithm, a probability that a restaurant has remaining stock for a house type with a demand gap, and collect historical data of the house type as a training sample.
The publishing unit 204 is configured to issue an offer to the hotel party when it is predicted that there is remaining stock in the restaurant party, where the offer is used to request that a room-type stock with a demand gap is added OTA.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that these are by way of example only, and that the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention, and these changes and modifications are within the scope of the invention.
Claims (4)
1. An OTA hotel inventory management method, comprising:
predicting the closing time length of each house type of the hotel on the OTA on different living days and whether a demand gap exists during closing;
the OTA hotel inventory management method predicts the room closing time of each room type of the hotel on the OTA on different living days through the following steps:
respectively establishing a house closing time length prediction model of different entering days aiming at each house type, wherein the house closing time length prediction model adopts an xgboost algorithm, collects historical data of the house type as a training sample, and selects one or more indexes from indexes used for reflecting one or more conditions of booking condition, browsed condition, price condition, tension condition or evaluation condition of the house type as input variables of the house closing time length prediction model;
the OTA hotel inventory management method predicts whether a demand gap exists when each room type of a hotel on the OTA is different to enter a closing room on a residence day through the following steps:
respectively establishing a prediction model of a closing inventory gap on different entering days for each house type, wherein the prediction model of the closing inventory gap adopts an xgboost algorithm, collects historical data of the house type as a training sample, and selects one or more indexes from indexes used for reflecting one or more conditions of booking conditions, browsed conditions, price conditions or tension conditions of the house type as input variables of the prediction model of the closing inventory gap;
for the house type with the demand gap, predicting whether a hotel party has residual stock;
the OTA hotel inventory management method predicts whether the inventory of a restaurant party is remained for the type of the room with the demand gap through the following steps:
calculating the probability of residual stock of the house type with the demand gap by a shop assistant by adopting a logistic regression algorithm, and collecting historical data of the house type as a training sample;
issuing an offer to the hotel party requesting an increase in the OTA of the house-type inventory for which there is a demand gap, upon predicting that there is inventory remaining at the hotel party.
2. The OTA hotel inventory management method of claim 1, wherein the OTA hotel inventory management method further comprises:
selecting the house type with the historical night amount larger than the night amount threshold value and related house history from all house types of all hotels on the OTA;
and predicting whether the historical night time is larger than a night time threshold value, and the closing time length of the house type related to the house history on different entering days and whether a demand gap exists during closing.
3. An OTA hotel inventory management system, comprising:
the first prediction unit is used for predicting the closing time length of each house type of the hotel on the OTA on different living days and whether a demand gap exists during closing;
the first prediction unit comprises a house closing time length prediction module and a house closing inventory notch prediction module;
the house closing time length prediction module is used for respectively establishing house closing time length prediction models of different entering days for each house type, wherein the house closing time length prediction models adopt an xgboost algorithm, historical data of the house types are collected to be used as training samples, and one or more indexes are selected from indexes used for reflecting one or more conditions of booking conditions, browsed conditions, price conditions, tension conditions or evaluation conditions of the house types to be used as input variables of the house closing time length prediction models;
the system comprises a house closing inventory gap prediction module, a house closing inventory gap prediction module and a database management module, wherein the house closing inventory gap prediction module is used for respectively establishing house closing inventory gap prediction models of different entering days for each house type, the house closing inventory gap prediction model adopts an xgboost algorithm, historical data of the house type are collected to be used as training samples, and one or more indexes are selected from indexes used for reflecting one or more conditions of booking conditions, browsed conditions, price conditions or tension conditions of the house type to be used as input variables of the house closing inventory gap prediction model;
the second prediction unit is used for predicting whether the hotel party has residual stock for the house type with the demand gap;
the second prediction unit is used for calculating the probability of residual stock of the type of the house with the requirement gap by adopting a logistic regression algorithm, and collecting historical data of the type of the house as a training sample;
and the issuing unit is used for issuing an offer to the hotel party when the fact that the inventory of the restaurant party is remained is predicted, wherein the offer is used for requesting to add the stock of the house type with the demand gap on the OTA.
4. The OTA hotel inventory management system of claim 3, wherein the OTA hotel inventory management system further comprises:
the selecting unit is used for selecting the house type with the historical night amount larger than the night amount threshold value and related house history from all house types of all hotels on the OTA;
the first prediction unit is further used for predicting the historical night time amount which is larger than the night time amount threshold value and whether a demand gap exists in the closing process or not when the house type related to the house history is different from the entering date.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611093188.4A CN106780173B (en) | 2016-12-01 | 2016-12-01 | OTA hotel inventory management method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611093188.4A CN106780173B (en) | 2016-12-01 | 2016-12-01 | OTA hotel inventory management method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106780173A CN106780173A (en) | 2017-05-31 |
| CN106780173B true CN106780173B (en) | 2021-02-23 |
Family
ID=58913828
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611093188.4A Active CN106780173B (en) | 2016-12-01 | 2016-12-01 | OTA hotel inventory management method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106780173B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109426887A (en) * | 2017-08-31 | 2019-03-05 | 阿里巴巴集团控股有限公司 | A kind of prediction technique and device of hotel inventory |
| CN107506877B (en) * | 2017-09-30 | 2020-12-01 | 携程计算机技术（上海）有限公司 | OTA platform-to-store no-room prediction method and system |
| CN107679674B (en) * | 2017-10-23 | 2021-06-11 | 携程计算机技术（上海）有限公司 | OTA platform overseas hotel room type service defect prediction method and system |
| CN107705191B (en) * | 2017-10-25 | 2021-03-26 | 携程计算机技术（上海）有限公司 | Order competition method and system of OTA platform |
| JP6864330B2 (en) * | 2017-11-20 | 2021-04-28 | オーブック・インコーポレイテッドＯＢＯＯＫ Ｉｎｃ． | Room inventory management system based on blockchain |
| CN111126652B (en) * | 2018-11-01 | 2023-05-16 | 爱满信息技术(上海)有限公司 | Hotel tail room intelligent prediction distribution system and method |
| CN111080417A (en) * | 2019-12-27 | 2020-04-28 | 携程计算机技术（上海）有限公司 | Processing method for improving booking smoothness rate, model training method and system |
| TWI786376B (en) | 2020-02-20 | 2022-12-11 | 緯創資通股份有限公司 | Demand forecasting method and demand forecasting apparatus |
| CN111445046A (en) * | 2020-03-18 | 2020-07-24 | 携程计算机技术（上海）有限公司 | Hotel reservation information processing method and system, electronic equipment and storage medium |
| CN112508528A (en) * | 2020-12-15 | 2021-03-16 | 携程计算机技术（上海）有限公司 | Method, system, device and medium for processing OTA website house closing |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104346654A (en) * | 2013-07-26 | 2015-02-11 | 三星Sds株式会社 | Demand forecast segmentation method, apparatus and method for adjusting demand forecast |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1755730A (en) * | 2004-10-01 | 2006-04-05 | 伺服网路科技股份有限公司 | On-line point number sale and distribution method |
| KR100698637B1 (en) * | 2006-04-07 | 2007-03-21 | 서용성 | A method for online selling and managing unitedly by franchise's participation of partial responsibility |
| JP5032692B1 (en) * | 2011-07-29 | 2012-09-26 | 楽天株式会社 | Reservation management device, reservation management method, reservation management program, and computer-readable recording medium storing the program |
| CA2968667A1 (en) * | 2014-11-26 | 2016-06-02 | Airbnb, Inc. | Demand prediction for time-expiring inventory |
-
2016
- 2016-12-01 CN CN201611093188.4A patent/CN106780173B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104346654A (en) * | 2013-07-26 | 2015-02-11 | 三星Sds株式会社 | Demand forecast segmentation method, apparatus and method for adjusting demand forecast |
Non-Patent Citations (1)
| Title |
|---|
| 需求预测综合模型及其与库存决定的集成研究;倪冬梅等;《管理科学学报》;20130930;第44-46页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106780173A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106780173B (en) | OTA hotel inventory management method and system | |
| Fradkin | Search, matching, and the role of digital marketplace design in enabling trade: Evidence from airbnb | |
| JP6543700B2 (en) | Determine host preference for accommodation listing | |
| KR102283033B1 (en) | Automated determination of booking availability for user sourced accommodations | |
| Ma et al. | Estimating warehouse rental price using machine learning techniques | |
| Espinoza et al. | Entrepreneurship and regional dynamics: The case of Chile | |
| CN110334289B (en) | Travel destination determining method and target user determining method | |
| Tse et al. | Analyzing the use of an advance booking curve in forecasting hotel reservations | |
| US20170293878A1 (en) | System and process for matching seniors and staffers with senior living communities | |
| CN106355526A (en) | Order-sending processing device, system and method | |
| CN109461030A (en) | A kind of O2O evaluation method and device based on block chain | |
| US20170178036A1 (en) | Unique accommodation search improvement founded in listing booking conversion | |
| CN111080417A (en) | Processing method for improving booking smoothness rate, model training method and system | |
| EP2681696A2 (en) | Project management system | |
| Arvanitidis et al. | Trust in peers or in the institution? A decomposition analysis of Airbnb listings’ pricing | |
| CN103309894A (en) | User attribute-based search realization method and system | |
| Grigolon et al. | Facet-based analysis of vacation planning processes: a binary mixed logit panel model | |
| CN110362728A (en) | Information-pushing method, device, equipment and storage medium based on big data analysis | |
| US20120226618A1 (en) | Project management system | |
| US11257011B2 (en) | Booked-blocked classifier | |
| US20140222700A1 (en) | Predictive pairing and/or matching systems, apparatus, and methods | |
| US20210004769A1 (en) | Automated event coordination | |
| Yan et al. | User recommendation with tensor factorization in social networks | |
| US20140143000A1 (en) | Project Management System | |
| TW202006617A (en) | Cloud self-service analysis platform and analysis method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |