Method and system for providing price availability, e.g., hotel services price availability - Patents.com

The method estimates fare class availability by comparing live bookable prices with calculated prices using statistical models and machine learning, addressing inefficiencies in existing systems by reducing data storage needs and server queries, thus enhancing speed and cost-effectiveness.

JP7587637B2Active Publication Date: 2024-11-20SKYSCANNER TECHNOLOGY LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
JP2023088699
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2013-12-11
Filing Date
2023-05-30
Publication Date
2024-11-20
Estimated Expiration
2034-09-18

AI Technical Summary

Technical Problem

Existing methods for determining fare class availability in travel services, such as airfares, are slow, expensive, and require significant data storage capacity and multiple server accesses, making them inefficient and costly.

Method used

A method and system that estimates fare class availability by comparing observable, live bookable prices with calculated prices using stored data, applying statistical models and machine learning techniques to infer availability without the need for extensive data storage or multiple server queries.

Benefits of technology

Reduces the time and energy required for determining fare class availability, eliminates the need for large data storage, and lowers operational costs while maintaining accuracy in price estimation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007587637000001
    Figure 0007587637000001
  • Figure 0007587637000002
    Figure 0007587637000002
  • Figure 0007587637000003
    Figure 0007587637000003
Patent Text Reader

Abstract

To provide a method and a server for providing fare availability that infers which fare classes are available.SOLUTION: A method for inferring fare class availability for a particular flight comprises: receiving an observable and live bookable price for the flight; comparing the bookable price to a set of prices calculated from stored data, such as valid itinerary data and fare data; using a statistical model for calculation of the set of prices; using statistical data obtained from a stored classifier corresponding to an identified group; determining which fare class was used at the bookable price, with a calculated price that most closely matches the bookable price; and providing inferred fare class availability to a computing device.SELECTED DRAWING: Figure 24
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The technical field relates to methods, servers, and computer program products for inferring what fare classes are available for travel or services, e.g., airfares, train fares, hotel prices, for goods, in fact, any type of goods or services whose prices are not fixed but instead variable and for which it is therefore useful to be able to provide estimates for availability for fare classes. [Background technology]

[0002] Global Distribution Systems (GDS) collect schedules from OAG (Official Airline Guide: see, for example, www.oag.com), fares from ATPCO, and fare class availability from the airlines, and store this information in their own systems. They periodically update this stored information from their sources. Performing an availability query against such a GDS system can be slow, expensive, and requires using energy to perform the search and transmit the results. Performing an availability query against many airlines is also slow, requiring access to multiple servers, and requires using energy to perform the search and transmit the results.

[0003] Storing fare class availability for all possible airports along all possible routes for a sufficiently long time horizon (eg, a year) would require quite significant data storage capacity.

[0004] FR2841018(B1) discloses a method for loading flight schedule changes, comprising the steps of receiving at least one set of flight changes, retrieving and storing individual changes in the set in a register for future scheduling, simulating reallocation of reservations related to the flight schedule changes by accessing both the records and the flight schedule database via a reservation distribution server, and finally updating the flight schedule database with reservation slack.

[0005] As an example, in the context of airfare flight pricing, flight pricing is traditionally obtained from a GDS (see Section C for a more detailed explanation of how GDS systems work, and more generally, how flight pricing works). Flight comparison services (e.g. Skyscanner) and some airlines pay a fee to obtain live bookable prices from a GDS. These prices are the actual bookable prices at which potential passengers can make a reservation. However, many potential passengers are likely to only be browsing and do not require a clearly accurate bookable price, but will instead be satisfied with a quote. As a result, it would be very useful if methods and systems could be devised to quote accurate prices, especially if it removed the dependency on accessing third-party resources (e.g. GDSs) that are not only expensive to access, but may also be unavailable to obtain the required information.

[0006] Thus, a method for obtaining estimates of fare class availability between any two airports over a period of time that is not as slow as multiple accesses to one or more remote servers to obtain the estimates, and that reduces the energy required to obtain the fare class availability estimates. It would be desirable to provide a method that is small in size and does not require the enormous data storage required to pre-store results for all possible queries.

[0007] According to a first aspect of the present invention, there is provided a method of inferring fare class availability for a particular flight, comprising the steps of: (a) receiving observable, live, bookable prices for its flights, where that pricing information is not accompanied by complete fare class information for those flights; and (b) comparing observable, live, bookable prices to prices calculated from stored data, e.g., sets of valid itinerary data, fare data, e.g., FROP data, and tax / surcharge data; (c) determining which fare class was used in the observable, live, bookable price by determining which fare class was used in the calculated price that most closely matches the observable, live, bookable price.

[0008] One advantage is that the method does not require the time or energy required to access one or more remote servers multiple times to actually set fare availability and calculate fare prices and pre-store the results for all possible queries. A further advantage is that the method does not require the significant data storage capacity required to pre-store the results for all possible queries. A computer performing the method operates in a novel manner to overcome the identified problems. The method may be one in which the stored data can be obtained from cached previously obtained live bookable prices.

[0009] The method may include a fare class, in which some or all of the applicable rules or restrictions that apply to a particular fare for a particular flight are defined by a fare class, which may include one or more of a class code, e.g., C (business full fare) or Y (economy full fare); a fare type, e.g., first class or economy; a category, e.g., eligibility, minimum stay, stopovers, or penalties.

[0010] The method may be such that an inferred fare class price is associated with each inferred fare class availability.

[0011] The method may include, in step (c), sending the inferred fare class availability to a server.

[0012] The method may be such that the server is an airline server.

[0013] The method may include step (b) including using the rules to analyze patterns in the set of prices calculated from the stored data.

[0014] The method may include computing the set of prices computed from the stored data in step (b) including building a statistical model from the historical prices, identifying missing candidate quotes, and pricing the candidate quotes based on the statistical model.

[0015] The method may be such that the input to the statistical model includes a reverse path equivalent.

[0016] According to a second aspect of the invention there is provided a server configured to perform the method of any of the first aspects of the invention.

[0017] According to a third aspect of the present invention there is provided a computer program product embodied on a non-transitory storage medium, the computer program product being arranged, when executed on a computer, to infer fare class availability for a particular flight, the computer program product, when executed on the computer, comprising: (a) receiving observable, live, bookable prices for its flights, where that pricing information is not accompanied by complete fare class information for those flights; and (b) comparing observable, live, bookable prices to prices calculated from stored data, e.g., sets of valid itinerary data, fare data, e.g., FROP data, and tax / surcharge data; (c) determining which fare class was used at an observable, live, bookable price by determining which fare class was used at a calculated price that most closely matches the observable, live, bookable price. A computer program product is provided that is configured to:

[0018] According to a fourth aspect of the present invention there is provided a method for inferring which offered fare classes are available, the method comprising the steps of: (i) receiving, by a computer server, a request for goods or services along with parameters defining those goods or services along with a price for those goods or services, e.g., airfare; (ii) at any time relative to step (i), configuring one or more processors to determine an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) configuring one or more processors to calculate estimates for requested prices for goods or services that satisfy the parameters; (iv) sending a request to a distribution system for a freight price; (v) receiving a distribution system freight price from the distribution system; (vi) comparing the calculated estimate for the requested price from step (iii) with the distribution system fare prices received in step (v) to estimate fare class availability; (vii) providing the inferred fare class availability to a computing device.

[0019] One advantage is that the method does not require the time or energy required to access one or more remote servers multiple times to actually set fare availability and calculate fare prices and pre-store the results for all possible queries. A further advantage is that the method does not require the significant data storage capacity required to pre-store the results for all possible queries. A computer performing the method operates in a novel manner to overcome the identified problems. It is noted that steps (ii) and (iii) in the method are not necessarily separate steps, but may instead be part of the same step.

[0020] The method may be such that the parameters defining the goods or services include one or more of the following: action type, e.g., airfare, hotel booking, train fare; date range; destination; origin; desired weather conditions; star rating; keywords; or any other user-defined preferences.

[0021] The method may determine the estimated price by inferring, deriving or predicting the estimated price.

[0022] The method further comprises the steps of: (a) obtaining historical price quotes from a computer data store; and (b) grouping past price quotes by category; and (c) deriving statistical data for each group; and (d) storing, on the computer, a classifier including the derived statistics for each group; and (e) using the stored classifier to identify the group to which the requested price corresponds.

[0023] The method may include in step (iii) calculating a set of quotes for the required prices over the specified date range using statistical data obtained from the stored classifiers corresponding to the identified groups.

[0024] The method may be one in which the distribution system is a global distribution system.

[0025] The method may be such that an inferred fare class price is associated with each inferred fare class availability.

[0026] The method may include in step (vii) sending the inferred fare class availability to a server.

[0027] The method may be such that the server is an airline server.

[0028] The method may be such that step (ii) comprises analysing patterns in the data set using the rules.

[0029] Step (ii) of the method may include a naive Bayes classifier machine learning approach to generate a probabilistic model of the prices, and use the model to predict unobserved prices.

[0030] The method may involve training a classifier using a set of observed prices and their corresponding features.

[0031] The method further comprises the steps of: the characteristics relate to the request and may include one or more of: day of week departure, length of stay, stay on Saturday, airline, time of day to travel, route, month.

[0032] The method may involve the classifier then predicting the price of an unobserved price given the set of features and the price that is most likely to have those features.

[0033] The method may derive features by training multiple models with different features and comparing the predictive accuracy of the different models.

[0034] The method may include, in step (ii), constructing a statistical model from historical prices, identifying missing candidate quotes, and pricing the candidate quotes based on the statistical model.

[0035] The method may include generating a price estimate for each candidate quote by extracting categorical feature values ​​from the quote, retrieving a trained classifier from a database for the extracted categories, extracting all feature values ​​from the candidate quote, and classifying the candidate quote by computing a Bayes posterior probability for each price range stored in the classifier, selecting the price range class with the highest Bayes posterior probability, and assigning the price class to the candidate quote.

[0036] The method may be one in which input to the statistical model includes a list of paths, a classifier categorization scheme, past estimates, and a set of supported features with weights.

[0037] The method may include filtering the past estimates by age.

[0038] The method may be such that the input to the statistical model includes a reverse path equivalent.

[0039] The method may include the step of including the cached freight prices in the set of price quotes.

[0040] The method may be such that the price is for a trip that is a one-way trip.

[0041] The method may be such that the price is for a trip that is a round trip.

[0042] The method may be such that the price includes an airfare price.

[0043] The method may be such that the price includes a rail freight price.

[0044] The method may be such that the price includes a car hire price.

[0045] In the method, the price may include a hotel price.

[0046] The method may be such that the request includes a flexible search request.

[0047] According to a fifth aspect of the present invention there is provided a server configured to perform the method of any of the fourth aspect of the present invention.

[0048] According to a sixth aspect of the present invention there is provided a computer program product embodied on a non-transitory storage medium, the computer program product being arranged, when executed on a computer, to infer which fare classes are available, the computer program product being arranged, when executed on the computer, to: (i) receiving a request for goods or services along with parameters defining those goods or services, such as a price for those goods or services, e.g., airfare; (ii) at any time relative to (i) above, configuring one or more processors to determine an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) configuring one or more processors to calculate estimates for requested prices for goods or services that meet the parameters; and (iv) sending the request to a distribution system for a freight price; (v) receiving the distribution system's freight price from the distribution system; and (vi) comparing the calculated estimates for the requested prices obtained from (iii) with the distribution system fare prices received in (v) to estimate fare class availability; (vii) providing the inferred fare class availability to a computing device.

[0049] According to a seventh aspect of the invention there is provided a computer terminal connected to a server of the second or fifth aspect of the invention, the computer terminal being arranged to send a request specifying fare class availability for goods or services, for example airfare, together with a request for parameters defining those goods or services, the computer further arranged to receive the output inferred fare class availability.

[0050] According to an eighth aspect of the present invention there is provided a good or service, for example an airline ticket, for which fare class availability is provided using a method according to the first or fourth aspect of the present invention, or using a server according to the second or fifth aspect of the present invention.

[0051] According to a ninth aspect of the present invention there is provided a web based price comparison website where an end user can interact with the web based price comparison website by providing a request for fare class availability and the web based price comparison website provides the request to a server which estimates fare class availability using a method according to the first or fourth aspect of the present invention.

[0052] The above and other aspects of the present invention will now be described, by way of example only, with reference to the following figures: [Brief description of the drawings]

[0053] [Figure 1] FIG. 1 is a diagram of an embodiment of a system for estimating prices from other prices with the goal of using the prices to increase the cache of observed fares. [Diagram 2] FIG. 13 illustrates an example of how a "difference" engine is used to derive fare class availability from live updates. [Diagram 3] FIG. 1 illustrates an example of a formula suitable for use in naive Bayes classification. [Figure 4] FIG. 13 illustrates an example where feature values ​​are represented by numbered regions and frequency information is represented by price ranges (buckets). [Diagram 5] FIG. 1 shows an example of the entire training process. [Figure 6] FIG. 13 illustrates an example where a naive Bayes classifier can be used to price each candidate quote placeholder. [Figure 7] FIG. 1 illustrates an example of the overall process for inferring missing estimates. [Figure 8] FIG. 1 illustrates an example of a server providing a quote service for fare availability and prices, and connections between input data sources and outputs to partners and users. [Figure 9] FIG. 1 illustrates an example of a user interface. [Figure 10] FIG. 1 illustrates an example of a user interface. [Figure 11] FIG. 1 illustrates an embodiment of a system for performing a search process. [Figure 12] FIG. 1 illustrates an example of a user interface. [Figure 13] FIG. 13 illustrates an example of a user interface output. [Figure 14] FIG. 13 illustrates an example of a user interface output. [Figure 15] FIG. 1 illustrates an example architecture for browsing searches. [Figure 16]FIG. 1 illustrates an example embodiment of an overall system architecture for a system that includes providing quote services for fare availability and pricing. [Figure 17] FIG. 2 is a diagram showing a GDS configuration embodiment. [Figure 18] FIG. 13 illustrates an example of a fare class record, fare categories, and a fare with respect to category data table. [Figure 19] FIG. 1 illustrates an embodiment of a process for providing fare pricing options. [Figure 20] FIG. 1 illustrates an embodiment of the relationships between the pricing engine, AVS / NAVS, Seamless (polling), Cache DB, and the airline host. [Figure 21] FIG. 1 illustrates an example system including an availability server. [Figure 22] 13 illustrates an example graphical display of confidence ranges for quoted prices using bars that span the price range from the lowest price to the highest price in the range in which a price should be found. Bars are provided for a range of outbound and return dates. [Diagram 23] FIG. 1 illustrates an example of major components in a system for deriving fare class availability and information flow within the system. [Figure 24] FIG. 1 illustrates an embodiment of a process for deriving fare class availability. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0054] One implementation of the present invention provides "model-based pricing," which is described below in Section A.

[0055] A. Model-Based Pricing Flight price estimates are provided using model-based data interpolation.

[0056] Background / Motivation The price of a ticket for an itinerary is complex to calculate; it is the result of calculations of possible routings, recognized fares, and the availability of these fares. These calculations are typically performed in a Global Distribution System (GDS), and performing queries against such systems can be slow and expensive.

[0057] Airlines that file fares with a GDS typically use what is known as a "schedule-based" pricing model. The important thing to note about such models is that round trips are scheduled on three or four days. 1 Departure date: 1 If you are returning to 1 ,R 1 ) is the price of two one-way tickets P(O 1 )+P(R 1 ) is not equal to P(O 1 ,R 1 )=P(O 1 )+P(R 1 ) is a model.

[0058] Two important consequences of scheduled pricing models are the limitations they impose on the reuse of observed prices and the space required to store them.

[0059] In the budget model, two "round trip" queries are made, one for the date O 1 R 1 On the other hand, the date O 2 R 2 For , the resulting price P(O 1 ), P(R 1 ), P(O 2 ), and P(R 2 ) to create up to four round trip fares (i.e., P(O 1 , R 1 ), P(O 1 , R 2 ), P(O 2 , R 1 ), P(O 2 , R 2)) and the four fares P(O 1 ), P(R 1 ), P(O 2 ), and P(R 2 If we run a query against the “schedule” model, we have only two fares: P(O 1 ,R 1 ) and P(O 2 ,R 2 ). Thus, in the budget model, 8 queries can be answered by 2 queries, whereas in the scheduled model, 2 queries can be answered by 2 queries.

[0060] Assuming flight prices are only available today and for the next 364 days, finding all possible one-way and round-trip fares on a route would require 365 round-trip queries. In the case of the schedule model, * It would require (365+1) / 2=66,795 round-trip queries, and 365 one-way queries (in each direction). Thus, to build a complete list of prices for routes, the scheduled model requires about 185 times more queries than the budget model.

[0061] If we had the results of all these queries, the storage required to store the fares would necessarily be about 185 times larger.

[0062] The above is merely an example for a route between two specified airports. In reality, there are about 5,000 civil airports in the world. Therefore, the number of possible trips between airports is about (5,000). 2 The aforementioned number 365 * Considering (365+1) / 2=66,795 possible round trip inquiries, this gives a total of (5,000) possible inquiries between all airports per year. 2* 365 * The result is (365+1) / 2, which is about 1012 The implicit amount of data storage is actually very large.

[0063] The greatly increased number of queries required to form a "full picture" of scheduled model routes combined with GDS costs means that it is impractical (i.e. inexpedient) to build such a cache, and the result is an incomplete cache, with "holes" where fares are not observed.

[0064] However, the price (returned by the GDS) is the result of applying fare rules and availability. As a result, there are patterns and consistencies in the observed fares. This embodiment uses these patterns to interpolate fares that are missing from the cache, using a model that reflects the fare rules that generated the price in the first place.

[0065] A system for estimating prices from other prices A system is provided for estimating prices from other prices with the objective of using the prices to augment a cache of fares that are observed (e.g., observed by a user requesting prices for live, bookable fares). A model embodiment is shown in Figure 1.

[0066] Estimators use statistics derived from estimates to predict other estimates. A very simple estimator might calculate that all missing fares are equal to the average of all observed fares. The level of prediction can be improved by making the model more sophisticated, taking into account the type of rules used to generate the fares in the first place.

[0067] For example, a common "rule" that dictates whether a fare is valid for an itinerary is whether the itinerary involves a Saturday night stay or not. Typically, the implication of such a rule is that dates that involve no Saturday night at the destination are much more expensive. The rule is designed to penalize business travelers, who usually want to return during the work week. Thus, a better model can be computed that includes two averages: the average of all observed fares with a Saturday night stay and the average of all observed fares without a Saturday night stay. The algorithm then predicts prices for previously unobserved date pairs by first computing whether there was a Saturday night between the date pairs and then returning the corresponding average. This results in increased accuracy of the estimates.

[0068] Similar improvements can be made to the model, for example by dividing travel prices by day of the week. Depending on the day of the week, it can vary (Mondays are typically more expensive than Wednesdays), so calculating and using an average that also takes into account the day of the week would also potentially improve the accuracy of the model.

[0069] A statistical model may thus be provided to predict fares, for example using statistical data and rules that may be human derived. Alternative ways of forming the statistical model are possible. In this approach, a machine learning approach called "Naive Bayes Classifier" is used to form a probabilistic model of fares, which is then used to predict unobserved fares. The classifier is trained using a set of observed prices and corresponding "features." Features could include day of the week departure, length of stay, Saturday stay, airline, etc., for the query that generated the fare. The classifier then predicts the price of the unobserved fare given the set of features and the most likely price having those features.

[0070] The statistical model for predicting prices can then be used to predict previously unobserved fares, or to predict observed fares (if one wishes to measure the accuracy of the model).

[0071] Main features and their calculation Features that may be used to train a classifier include: Stay on Saturday Length of stay Day of the week Time to go on a trip Airlines Website ·route Date-day and year-month level ·month

[0072] These features are derived by analyzing all fares listed in FROP (Fare Rules Output Product, from ATPCO (Airline Tariff Publishers)) and determining which rules apply to the largest number of fares. For example, 54% of fares have a minimum stay rule, and 31% of all fares have a minimum stay rule that requires a Saturday night at the destination. Thus, the features may be useful features for classifying based on price. A more accurate estimate of the usefulness of the features may be derived by training multiple models with different features and comparing the predictive accuracy of the different models.

[0073] profit Estimate missing flight price data Use data-generated features to create better predictive models

[0074] Potential markets and / or application areas Models may be used in the flight industry to estimate previously unobserved prices, to display prices to a user, or to determine whether it is worth querying to return a price (e.g., it may be a good value).

[0075] This model has the disadvantage of not capturing prices (e.g., inquiry costs) in other areas, but it does have the advantage of generating prices through structured means (e.g., pricing rules). This includes travel products such as hotels, trains, and car rentals, as well as other products with structured pricing, such as insurance.

[0076] Other Applications Consumer price comparison websites. Companies attempt to price competitively by comparing their prices with those of their competitors.

[0077] Model-Based Pricing - Inferring the Missing Estimates Inference of missing estimates may be done in two stages: training and estimation. ·training ○Build statistical models from past prices ·estimate Identify missing quote candidates Price candidate quotes based on statistical models

[0078] In an embodiment, the statistical model is a set of naive Bayes classifiers that are trained periodically (e.g., daily) using historical prices. Missing quotes are inferred based on the route operator's data and priced using the classifiers.

[0079] Statistical model implementation Naive Bayes classification is based on the dependency between the probability that an object described by a set of features belongs to a class and the probability that an object from this class is characterized by a set of features, see equations (1) and (2) in Figure 3.

[0080] In FIG. 3, equations (1) and (2) can be simplified to equation (3). In equations (1) to (3), C is the class into which the input object can be classified (for example, in this case, the price range, e.g. 75-125 EUR). F i is a feature value (for example, it is Friday, the Air France airline website, etc.).

[0081] The classification can be expressed using equation (4) in Figure 3.

[0082] For a given set of feature values ​​for which we wish to find a class (price range), we choose the feature value that occurs most frequently and therefore has the greatest probability of the class characterizing the input set.

[0083] training The intent of classifier training is to calculate how frequently feature values ​​occur for observed price ranges. For example, in the process shown in Figure 4, feature values ​​are represented by numbered regions and price ranges (buckets) represent frequency information (e.g., feature value quantities range from 75-125 EUR, 126-175 EUR, etc.).

[0084] The following elements are input for the training process example: Route list Classifier category classification method Past quotes Supported feature set with weights

[0085] The list of routes sets the boundaries for training. A range of classifiers is defined, and categories are represented by a set of feature values. The category we empirically chose consists of city-level routes and airlines (e.g. London-Paris, Air France), but other options are possible (e.g. city-level routes without airlines). There is always one classifier per category. We train the classifier using only past estimates matching that category (e.g. only estimates for the London-Paris route (operated by Air France)).

[0086] The first step of the training process is to take a certain number of recent past quotes from the database with some maximum age. Evaluate the categories for every quote. Group the quotes by category. Each group becomes the training set for one classifier.

[0087] The training set can be expanded with estimates for the reverse route equivalent. Example: For a historical estimate on the route London-Stansted-Rome-Ciampino (Air France), the reverse route equivalent would be Rome-Ciampino-London-Stansted (the airline and other details, e.g. price, are the same). Include estimates for the reverse route in the training, scaled by a weight that adjusts their importance.

[0088] Then, for each group, we build a classifier with the following statistics extracted from the estimates: Weighted number of estimates including each distinct feature value Weighted number of quotes that include each distinct feature value per price range of some granularity · Many observed price ranges, estimated price ranges, and all estimates

[0089] The weights applied to the various estimates can be varied across features to highlight the greater influence of some of them (for example, the inclusion of a weekday in a trip may be more important than the length of the stay).

[0090] The process ends with one classifier for each training set (e.g., one classifier for the London-Paris route (operated by Air France)) containing statistical information. The models are stored for future use. The classifiers can be trained again, for example periodically or when many past estimates change.

[0091] A classifier may be formed only if a certain number of estimates matching that category are available.

[0092] An example of the entire training process is shown in Figure 5.

[0093] In addition to the parameters listed above, the training process can be parameterized by the following: Price range size per route (e.g. 1EUR, 5EUR, 10EUR) Weighting applied to reverse paths ·Weighting function to apply based on estimated age

[0094] Estimate the missing estimates In an embodiment, a user searches for one-way or round-trip flights for a route and date range. The system responds to the query with a set of cheapest quotes from the database, at most one for each date (or date pair in the case of round-trip flights). The response may not contain quotes for some dates (date pairs). Missing quotes are ignored. Inferring the estimated probability can be done using a trained naive Bayes classifier.

[0095] The following elements may be input to the inference process: ·route A list of dates (or date pairs) with missing best price quote information

[0096] Since the classifier can only be used to price existing quotes, candidate quotes must be evaluated first. Candidates are all applicable airport pairs, dates (date pairs), airlines, and agency combinations available for the query route. The generation of the list is based on the process inputs and the database of the route operator.

[0097] We then price each candidate quote placeholder using a naive Bayes classifier, see for example Figure 6.

[0098] A price estimate may be made for each candidate quote by the following steps. Extract categorical feature values ​​from the estimates. Extract a classifier trained on the extracted categories from the database. Extract all the features from the candidate estimates. - Classify candidate quotes by: Calculate the Bayesian posterior probability for each price range stored in the classifier Select the price range class with the highest Bayesian posterior probability · Assign price classes to potential quotes

[0099] If no classifier exists for a candidate quote, no price guess is made and the date (date pair) slot is left empty.

[0100] The process can result in multiple priced quote candidates per date (date pair). For this reason, the next step is to select a single quote for each date (date pair). The selection can be based on price. For example, select the cheapest quote per date (date pair).

[0101] The end result of the process is, at most, one estimate per input date (date pair). The inferred estimates can be included in the response returned to the user. An example of the complete process for inferring missing estimates is shown in Figure 7.

[0102] In addition to the parameters already listed, the inference process can be parameterized by: The minimum Bayesian posterior probability required to accept a classification result · Maximum number of route operators involved in candidate generation Random variation added to Bayesian posterior probabilities to avoid ties

[0103] A quote method that gives you price confidence An estimation method MAY provide some confidence in the price, e.g., they are 99% confident that it is within $5, within $1000, etc. Thus an estimation method MAY provide: - Calculation and return of these beliefs. These beliefs are used to determine whether to display the price to the user or to show the user a likely price range (e.g., error bars).

[0104] Many methods for estimating prices can return a measure of confidence in the price. For example, in the case of a naive Bayesian classifier, the class selected is the one with the highest probability, and this probability can be used as an indicator of confidence in the estimate. The closer the probability is to 1, the more confident we are that the predicted price is accurate.

[0105] The resulting confidence may be used in a number of ways. If the confidence in the estimate is very low, the following decisions may be made: ○ Don't provide estimates because they are not useful to users ○ Trigger a live fetch of the price so that the exact value can be shown to the user. The confidence can be shown to the user. This can be shown in a number of ways. A bar showing the range of possible values ​​for the price (see, for example, Figure 22). Reduce the visibility / intensity of the displayed price (e.g. by using coloring, shading, or dotted lines).

[0106] The confidence range for the quoted price spans the price range from the lowest price to the highest price within which the result of the estimation methodology leads to confidence that a price will be found. The confidence range for the quoted price may be shown graphically using a bar. A bar may be displayed for each date in the date range. A bar may be displayed for each date in the date range, for outbound and inbound journeys. The bars may be color coded, for example, to indicate to the booking function that the bar corresponds to a recent price that is likely to be stable for a price range derived from multiple tickets sharing a departure date and for a price range quoted from a similar itinerary. One bar may be provided for a date. Multiple bars may be provided for a date. An example is shown in FIG. 22.

[0107] B. Derived Availability Engine Background / Motivation The motivation is very similar to that in model-based pricing implementations.

[0108] The implication of a "scheduled" pricing model is that two single quotes cannot be used to form a round trip quote, so a query must be run for all possible date pairs, which is prohibitive due to the large number of date pairs, due to the cost of obtaining this data and the space required to store it. However, the round trip price is the result of combining the available fares available for the outbound and return legs, together with some rules. Thus, if the availability and rules are known, it is possible to decompose a round trip query into "reusable" legs and use them to construct a new "round trip" price, in a similar (but different) way to that used for airlines with budget pricing models.

[0109] Example of a derived availability engine The Global Distribution System collects schedules from OAG (Official Airline Guide: see, for example, www.oag.com), fares from ATPCO, and fare class availability from the airlines, stores this information in its own system, and periodically updates this stored information from those sources. When a GDS is queried for a quote for a route and dates, the GDS performs the following steps: 1. Determine valid itineraries from the schedule. 2. Calculate valid fares for the itinerary (some GDSs do not have their own pricing engine and use third party engines, such as those provided by SITA (see for example http: / / www.sita.aero / ), to ensure that the correct price is applied for each travel itinerary). 3. Find the availability of these fares. 4. Add the correct taxes and surcharges.

[0110] Step 1 may be implemented in a route service. Specifically, step 1 may be implemented using timetables and rules about what can be sold. Upon user query, valid itineraries for which tickets can be purchased can be identified.

[0111] Airlines send fares to ATPCO, and ATPCO provides offers for these data. However, it would be a huge effort to convert these raw fares and rules into a system that would determine the correct fare for a given inquiry. Fortunately, ATPCO provides combined fare and rule data in a data feed called FROP (Freight Rules Output Product). It sends the FROP data in a fixed length record file that contains rule conditions and fare information organized for major categories. So step 2 can be done with the FROP data. It can also be achieved by combining the fare rules and prices from the airlines with the pricing engine.

[0112] ATPCO also provides a feed of itinerary and service charge taxes. As a tax and surcharge engine, we can build an engine that takes the data in these feeds and uses them to calculate the total price a customer would pay for a fare, thus emulating step 4.

[0113] The only remaining item for low cost calculation of GDS fares is availability - step 3. Availability can be derived by comparing the results of live updates (the final result of steps 1-4) with the calculation of steps 1, 2 and 4 outlined above, i.e. by combining the route services, fares obtained from FROP, and the taxes and surcharges engine, all possible fares can be calculated and the observed fares from live updates can be used to infer which of the possible fares were used and therefore which of the calculated fares were available. The diagram in Figure 2 shows how such a "difference" engine can be used to derive fare class availability from live updates.

[0114] When available fares are known for both the outbound and return legs of a round-trip flight, round-trip price calculations can be made even if a particular date pair in the round-trip has not been observed. This method of modeling quotes is known as Single Leg Combinable (SLC). SLC is similar to the "budget" pricing model, except that in budget cases the outbound and return legs of a round-trip are also valid as one-way tickets. In budget cases, constructing a return trip is simply a matter of adding the return and outbound leg prices (the SLC construction is slightly more complicated due to the need to apply availability after finding valid fares for the date pairs).

[0115] If the price of a fare increases but the availability remains the same, it is possible to use the derived availability engine to calculate the new price even if it has not been observed.

[0116] Providing derived availability in this way has an additional benefit compared to using a GDS: doing a pure calculation in a GDS does not guarantee that the fare can be "found" on an airline or OTA website, whereas an observed fare can, by definition, be found. Also, a pure calculation does not take into account surcharges, markups, discounts, and private fares (which may be built into the observed fare). .

[0117] Note that fare prices do not necessarily change; there may be many different fares for a particular flight, and the availability of each fare may change. Because revenue management software may continually change which fares are available, it is possible for a cheaper ticket to disappear as revenue management cancels it to increase revenue.

[0118] Further Examples of Derived Availability Engines A derived availability engine may use fare, tax and surcharge data from a central registry (e.g., ATPCO FROP) to evaluate availability information (class of service) for a given set of observed quotes.

[0119] An example of the main components in the system and information flows is shown in FIG.

[0120] Observed quotes can come from OTAs, airline websites, APIs, and other services. Each quote can include airport-level routing, airline, number of flights, price (fare and / or tax) and other details.

[0121] An example of a fare source is the ATPCO product (called FROP). The FROP is available for application fares and contains fares in a flat, tabular format. Information contained in each FROP record is, for example, city-level routing, airline, class of service, price, and constraints (e.g., advance purchase specifications) that describe when the fare can be applied. The FROP is constructed from data published by airlines.

[0122] Tax and surcharge data can be obtained from public sources (e.g., airport or government websites) and the data is provided by ATPCO as a pre-paid product. It covers most government taxes, airport fees, fuel charges, and other components added to the fare. The final price paid by the passenger is the fare price plus the tax and surcharge amounts.

[0123] Using the components listed, an attempt can be made to derive an estimated availability (service class) for each observed resource by applying the process described below.

[0124] The first step in the process is to evaluate the set of observed quotes that are of interest (e.g., by some search criteria, e.g., route, travel date, passenger information, etc.). Then, as FROP records, we need to find quotes that match the observed search query.

[0125] With both sets of prices, we have to convert them into a comparable form. The next step depends on whether the observations contain only untaxed or only taxed prices. In the first case, the observed fares can be compared directly with the FROP. In the other cases, we have to calculate taxes and surcharges and add these to all FROP fares, taking into account the query conditions (e.g. number of passenger types, point of sale, etc.).

[0126] By putting the observed quote and the FROP in the same format (taxed or nontaxed), we can compare the quotes and find matches. We assume that quotes and FROPs are equivalent if the prices are the same.

[0127] Depending on the form of the fare from the authority (in this case FROP), the matching process may be further Further steps may be required to construct round trip fares from multiple one-way records and match them to observed round trip quotes.

[0128] At this point, the link between the FROP record and the observed quotes has been established. Then, for each observed quote, it is possible to derive the fare class availability by applying the service classes from the equivalent FROP record.

[0129] An embodiment of a process for deriving fare class availability is shown in FIG.

[0130] profit Derives availability rather than having to query a GDS or airline reservation system to find availability. · Calculate new prices based on estimated availability rather than true availability · Quote new prices based on availability and fare rules without having any observed instances of the new prices.

[0131] Other potential applications Provide airlines with pricing for their own products, as they currently obtain this pricing from the GDSs at high cost.

[0132] The company is interested in finding flight data prices at a fairly low cost.

[0133] Example of an architecture overview A server (which may be real or virtual) is provided on which fare availability and price quoting services are provided, which may provide associated web sites, smartphone apps, business-to-business services, connections with traditional airlines (e.g. British Airways, Qantas, KLM, etc.), connections with budget airlines (e.g. Ryanair, easyJet, Germanwings, etc.) and travel booking web sites (e.g. lastminute.com, Opodo, Thomson, etc.).

[0134] Travel booking web sites and airlines may provide input data to a server that provides fare availability and price quote services. The server may provide output via a public application programming interface (API) to, for example, partner companies. The server may provide output to search web sites, search mobile web sites, and apps on smartphones for users. An example of the connections between the server and input data sources, partners, and output to users is shown in FIG. 8.

[0135] In a search on the server, the user may specify a departure location (which may be, for example, an individual airport, a city with multiple airports, or a country with multiple airports). In a search on the server, the user may specify a destination location (which may be, for example, an individual airport, a city with multiple airports, or a country with multiple airports). In a search on the server, the user may specify a departure date or a range of departure dates (which may include, for example, a date range, or a calendar month, or an unspecified date (e.g., any date in the next year)). In a search on the server, the user may specify a return date or a range of return dates (which may include, for example, a date range, or a calendar month, or an unspecified date (e.g., any date in the next year)). In a search on the server, the user may specify a one-way ticket or a round-trip ticket. The user may specify the number of passengers in a particular age range. The user may specify a class of ticket (e.g., economy, premium economy, business class, or far A parallel search may be performed for hotels. A parallel search may be performed for car rentals. A preference for direct flights may be indicated. A user interface example is shown in FIG. 9. A further user interface example is shown in FIG. 12. This is an example of a flexible search because the departure indicates that any airport in London (UK) can be used and the destination indicates that any airport in Australia is allowed.

[0136] The output search results may include a list of flights that meet the search criteria. A graphical indicator (e.g., a slider bar) may be provided to define a departure time range for outbound flights. A graphical indicator (e.g., a slider bar) may be provided to define a departure time range for round trip flights. Selectable tabs may be provided to arrange the flights in order of increasing cost. Selectable tabs may be provided to arrange the flights in order of increasing travel time. Selectable tabs may be provided to arrange the flights in alphabetical order by airline name. An example user interface is shown in FIG. 10.

[0137] In an embodiment of a user search, flights between an origin and a destination (which may be a one-way flight or a round trip flight) are requested for outbound flight dates and, if applicable, for round trip flight dates. In an embodiment of the search process, the first step is to select a partner search web site that shows the output. The second step is to identify relevant prices already cached. The third step is to identify and update relevant prices that are not available. The fourth step is to cleanse and store the new data (e.g., taxes and fees may be removed and who the actual flight operator is may be identified). The fifth step is to return the prices to the user. An embodiment of a system for performing the search process is shown in FIG. 11.

[0138] In a flexible search, the output data may provide a list of destination airports that meet the flexible search criteria (e.g., destination is any airport in Australia, departing from any airport in London, UK). The list may be arranged in ascending order of price. An example of user interface output is shown in FIG. 13. Once a particular airport is selected, the lowest price for the outbound leg of the trip may be shown as a function of outbound departure date, e.g., for selectable calendar months, and the lowest price for the return leg of the trip may be shown as a function of return departure date, e.g., for selectable calendar months, for the selected departure and destination cities or airports. An example of user interface output is shown in FIG. 14.

[0139] A browsing service may provide a browsing search. A browsing service may be provided to users of a smartphone app or to users of a search web site. The browsing service may access the browsing price data. The browsing price data may include data scraped from web site content. The browsing service may connect to a public application programming interface (API). The browsing service may not be connected to airline servers or to the servers of other travel information service providers. An example architecture for browsing search is shown in Figure 15.

[0140] An example of an overall system architecture for a system that includes providing quote services for fare availability and pricing is shown in Figure 16.

[0141] The following sections provide background information on how air fares are calculated and the role of the GDS.

[0142] C. Introduction to the Global Distribution System A Global Distribution System (GDS) may be defined as a system that includes availability, pricing, and related services for airlines through which reservations and tickets can be issued. The GDS makes these functions available to applicant travel agents, booking engines, airlines, and other applicants. An example GDS configuration is shown in FIG. 17. A more general term is "distribution system." This may include the ability to make reservations for modes of transportation other than air transportation (e.g., trains, etc.).

[0143] Route planning Historically, routing was added by the GDS as airline networks grew. When the number of potential possible connections grew exponentially (SOF-NYC has many connections, e.g. SOF-IST-NYC, SOF-FRA-NYC, SOF-MOW-CPH-NYC, etc.), the GDS added this layer of "connection" construction to reduce the number of different priced routes so they don't have to deal with many fares and rules. They did it in a very obtuse way - static route and connection tables. This way they tell them to fly between zones (defined by ATPCO-US-EUR, EUR-APAC for simplicity of discussion), essentially route finding cities between SOF-NYC, including joins in the potentially wrong direction (e.g. IST), but rejecting a large number of cities as possible connections. For example, the flight SOF-PRG-CPH-NYC may be considered too complex and potentially expensive, and so is dropped at the route finding stage.

[0144] Connection Building After the routing is set, connection building starts looking at airlines, flight numbers and connection restrictions (contained in the Minimum Connection Time (MCT) table). The MCT (distributed by OAG) is a complex document that specifies the general rules on how flights / airlines can connect in terms of time. For example, if a long haul US-EU connects to EU-EU at the LHR, a 120 minute MCT applies. The general rules are simple to interpret, but important. If someone tries to book outside the MCT parameters, a manual override is required.

[0145] What is interesting are the MCT exceptions (around 200,000 or so). This goes so far as to say that DL flight #003 connecting to AF flight #004 might require 55 minutes. However, if DL flight #003 arrives at gate 19 at CDG and connecting flight AF009 departs from gate 49, then 65 minutes as MCT applies, etc. Of all the exceptions, around 70,000 are for Air France (AF) and Paris-Charles de Gaulle (CDG), which is why many passengers consider it the worst place to fly.

[0146] fare An example of a fare is a pricing contract, which specifies the price (base fare) for the services provided by an airline. The price is specified by class of service, class type (Y, Q, J, M, H, etc.), and type of trip (one-way, round trip, multi-city, round-the-world (RTW)).

[0147] In an embodiment, each fare has an associated number of rules and restrictions that apply to it, so think of fares as contracts 20-30 pages long, not just Excel rows (spreadsheets or CSV files). Instead of thinking of fares as rows in a big Excel spreadsheet, with columns indicating what features the fare has, you should instead think of fares as complex things in their own right, with many parameters and restrictions, and thus more like a contract. Fares define who, when and how fares are priced based on fare-passenger type. It specifies when the fare can be used, when the fare is applicable (NVB, NVA - not valid before, not valid after), and many other restrictions. A fare can be a public fare (accessible to any reseller) or a private fare (accessible to a designated reseller / vendor).

[0148] Fares are given by airlines. They are set at a general level for the market and are limited by travel and ticketing dates. Naturally, fares represent price points.

[0149] Fare upload and distribution ATPCO is the trade association that handles fares and fare distribution in the industry. There are a few airlines that only distribute to SITA (Bulgaria Air is an example).

[0150] Scheduling Historically, ATPCO has released fares on a specific schedule, but due to airline demand, the organization has begun releasing fares hourly, allowing airlines to ensure rapid response to market changes that are not only amenable to availability control.

[0151] SITA publishes its fares four times a day during the week and three times a day on weekends, and has fares that exist within the GDS (GDS specific fares or negotiated fares), which are entirely under the control of the GDS and do not follow a general distribution schedule.

[0152] Fare rules, restrictions and categories Provides fare data, fare class data, and fare restriction data.

[0153] How does pricing work (at a very high level)? Rules are calculated by the pricing engine in a very strict order of rule records, tables and sub-tables. If a specific rule record is not found, there is a general rule that governs them all.

[0154] Restricted Data -The fare class tells the engine that there are rules for a particular fare. -Category controls tell the engine what type of restrictions exist. - Category data tables are the last thing engines read to get relevant information.

[0155] For example, a fare class tells the engine that it has a rule attached to it. The rule is category 15 (sales restrictions), which after consulting the category data table, pulls out additional information about it (what type it is, what the rules and restrictions are).

[0156] An example of a fare class record, fare categories, and a fare to category data table is shown in FIG.

[0157] Fare Display vs. Fare Quote and Process Many industry professionals use the terms freight indication and freight quote interchangeably, but this is unacceptable to the skilled artisan.

[0158] Fare display A comprehensive, integrated concept of market fares and associated rules that enables agents to: -Compare fare rules (NB! not prices) -Inform passengers of the contract period (fare rules)

[0159] Fare Quote The final price after applying the contract period. This captures most of what we want to give our passengers. -All taxes -All additional charges - All rules related to it -Final price

[0160] An embodiment of a process for providing fare pricing options is shown in FIG.

[0161] Availability Knowledge of availability is very important in the industry. Availability allows airlines to control their headroom (seats) and also adds another way of managing how profitable a route is, by opening and closing availability buckets. This is managed by Revenue Management Systems (RM).

[0162] Some RMs are highly capable of managing complex rules, while others have a high degree of complexity since most of the availability rules are formulated by an internal corps of statisticians and mathematicians.

[0163] When one wants to check if a seat / class of service is available or not, there are three main ways to receive the availability:

[0164] i) NAVS / AVS NAVS=Simple Numeric Availability per Seat Class. Example: Y9H3M0L8. AVS = Availability Status - this is the flight open / closed, open for waitlist, etc.

[0165] ii) Polling CXRs (Airlines) can use NAVS / AVS as an indication if a better connection falls off. These CXRs use revenue management rules and must be polled to get actual availability and all availability information.

[0166] Some airlines and / or flights require polling, and the instruction for this is MR (Market Restriction). When the GDS requests an AVA, if the tag is MR, the GDS must poll for the correct availability. Unfortunately, NAVS / AVS does not show these tags, so when the GDS uses NAVS / AVS for tagged flights, it may show incorrect availability.

[0167] iii) Cache DB To reduce polling and NAVS / AVS, pricing engines have developed complex cache DBs. Although they are not accurate, they do achieve some of the goal of reducing calls when storing information.

[0168] An example of the relationships between the pricing engine, AVS / NAVS, Seamless (polling), Cache DB, and the airline host is shown in FIG.

[0169] Availability Server Example A caching mechanism may be used by the online channels as well as by the offline traffic. The availability server may be able to update the cached data based on rules (two-way polling) and alert on even the slightest discrepancies to reservations. An embodiment of a system including an availability server is shown in FIG. 21.

[0170] Note Implementing the steps described above can be accomplished using standard, well-known programming techniques. The novelty of the above-described embodiments lies not in any particular programming technique, but in using the described steps to achieve the described results. Software programming code embodying or forming part of the present invention is typically stored in persistent, non-transitory storage. In a client / server environment, such software programming code may be stored using storage associated with the server. The software programming code may be embodied on any of a variety of known media for use with data processing systems (e.g., diskettes, or hard drives, or CD-ROMs). The code may be distributed on such media, or distributed to users, from the memory or storage of one computer system over some type of network to other computer systems for use by users of such other systems. The techniques and methods for embodying software program code on physical media and / or distributing software code over a network are well known and will not be described further herein.

[0171] It will be understood that each element of the description, and combinations of elements in the description, can be implemented by general purpose and / or special purpose hardware-based systems that perform the functions or steps specified, or by a combination of general purpose and / or special purpose hardware and computer instructions.

[0172] These program instructions may be provided to a processor to form an apparatus, such that the instructions, executed on the processor, form means for performing the functions specified in the description. The computer program instructions may be executed by a processor to cause the processor to perform a series of operational steps to form a computer-implemented process, such that the instructions, when executed on the processor, provide the steps for performing the functions specified in the description. Accordingly, the figures support combinations of means for performing the specified functions, combinations of steps for performing the specified functions, and program instruction means for performing the specified functions.

[0173] It will be understood that the foregoing arrangements are merely illustrative of the application of the principles of the present invention. Numerous modifications and alternative arrangements can be devised without departing from the spirit and scope of the invention. While the invention has been shown and described above with particularity and detail in the drawings and fully described above with respect to what are presently considered to be the most practical and preferred embodiments of the invention, it will be apparent to those skilled in the art that numerous modifications can be made without departing from the principles and concepts of the invention as set forth herein.

[0174] concept Several concepts (hereinafter referred to as Concepts "A-J") are included in this disclosure. The following may be useful in defining these concepts. Aspects of the concepts may be combined.

[0175] A. How to provide a price quote 1. A method of providing a price quote, comprising: (i) receiving, by a computer server, a request for goods or services along with parameters defining those goods or services along with a price for those goods or services, e.g., airfare; (ii) at any time relative to step (i), configuring one or more processors to determine an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) configuring one or more processors to calculate estimates for requested prices for goods or services that satisfy the parameters; (iv) sending the price quote to an end user computing device, such as a personal computer, a smartphone or a tablet.

[0176] The foregoing may further include any of the following, either alone or in combination: In the present method, the parameters defining those goods or services include one or more of the following: action type, e.g. airfare, hotel booking, train fare; date range; destination; origin; desired weather conditions; star rating; keywords; any other user-defined preferences. · Determining price quotes by inferring, deriving or predicting price quotes. Step (ii) includes (a) obtaining historical price quotes from a computer data store; (b) grouping the historical price quotes by category; (c) deriving statistics for each group; (d) storing on the computer a classifier that includes the derived statistics for each group; and (e) using the stored classifier to identify the group to which the requested price corresponds. · Step (iii) involves calculating a set of quotes for the required prices over the specified date range using the statistical data obtained from the stored classifiers corresponding to the identified groups. · Step (ii) involves analysing patterns in the data set using the rules. · Step (ii) involves a naive Bayes classifier machine learning approach to generate a probabilistic model of prices and use that model to predict unobserved prices. · Train a classifier using the observed prices and their corresponding feature sets. · The characteristics relate to the request and include one or more of: day of week departure, length of stay, stay on Saturday, airline, time to travel, route, month. The classifier then predicts the price of an unobserved price, given a set of features and the most likely price to have those features. Derive features by training multiple models with different features and comparing the predictive accuracy of the different models. · Step (ii) includes building a statistical model from historical prices, identifying missing quote candidates, and pricing the quote candidates based on the statistical model. A price estimate is generated for each candidate quote by extracting categorical feature values ​​from the quote, retrieving a trained classifier from a database for the extracted categories, extracting all feature values ​​from the candidate quote, and classifying the candidate quote by computing a Bayes posterior probability for each price range stored in the classifier, selecting the price range class with the highest Bayes posterior probability, and assigning the price class to the candidate quote. The inputs to the statistical model are a list of routes, a classifier category classification scheme, and past estimates. It contains a set of supported features along with their weights. Filter past quotes by age. · Input to the statistical model includes reverse path equivalents. The method includes the step of including the cached freight prices in a set of price quotes. · The set of price quotes is constructed after step (i) without querying the distribution system. ·Prices are for trips which are one-way journeys. ·Prices are for trips which are round trips. The price includes airfare. -Price includes rail freight. The price includes car rental. The price includes hotel rates. · The request includes a flexible search request. ·The end result of the method is at most one quote per input date or pair of input dates.·The method includes the steps of: (A) configuring one or more processors to determine, at any point in time relative to step (i), a confidence range of estimated prices from an incomplete historical price data set by analyzing patterns in the data set; and (B) configuring one or more processors to calculate a confidence range of estimated prices for requested prices for goods or services that satisfy the parameters. The method further includes (C) providing the fare price quote along with the confidence range to an end user computing device, such as a personal computer, a smart phone, or a tablet. The method further includes using the belief range to determine whether to display a price to the user or provide the user with a likely price range. Show likely price ranges as error bars. -A date range includes only one departure date. -The date range includes only one return date. The method includes providing a set of best round trip fare price quotes for a specified departure date range and a specified return date range, where step (i) includes receiving a request for round trip fare prices for a specified departure date range and a specified return date range for travel from an origin to a destination, and step (iv) includes providing the set of best round trip fare price quotes for the specified departure date range and the specified return date range. A set of best round trip fare price quotes is provided in a bar chart format. The estimation process is parameterized by one or more of the following: the minimum Bayes posterior probability required to accept the classification result, the maximum number of path operators involved in generating the candidates, or a random variation added to the Bayes posterior probability to avoid ties. The method is performed on a server where a quote service for fare availability and prices is provided.

[0177] A server configured to provide price quotes, (i) receiving a request for goods or services along with parameters defining those goods or services, such as a price for those goods or services, the parameters including one or more of an action type, e.g., airfare, hotel booking, train fare; date range; destination; origin; desired weather conditions; star rating; keywords; any other user-defined preferences; (ii) at any point in time relative to (i) above, determining an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) calculating an estimate for the price requested for goods or services that meet the parameters; and (iv) providing price quotes; and a server configured to do so is provided.

[0178] The server may further include any of the following, either alone or in combination: (ii) the server is configured to obtain historical prices from a computer data store, group the historical prices by category, derive statistics for each group, store a classifier for each group that includes the derived statistics, and identify the group using the stored classifier to which the requested price corresponds. The server is configured to calculate a set of quotes for the requested price over a specified date range using statistical data obtained from the stored classifiers corresponding to the identified groups. The server is further configured to perform a method according to any of the aspects of Concept A.

[0179] A computer program product embodied on a non-transitory storage medium, the computer program product being arranged to provide a price quote when executed on a computer, the computer program product being arranged to: (i) receiving a request for goods or services along with parameters defining those goods or services, such as a price for those goods or services, the parameters including one or more of an action type, e.g., airfare, hotel booking, train fare; date range; destination; origin; desired weather conditions; star rating; keywords; any other user-defined preferences; (ii) at any point in time relative to (i) above, determining an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) calculating an estimate for the price requested for goods or services that meet the parameters; and (iv) providing a price quotation; and a computer program product arranged to do so is provided.

[0180] A computer program product may be arranged to carry out the method of any aspect of Concept A.

[0181] There is provided a computer terminal connected to a server of any of the servers of Concept A, the computer terminal arranged to send a request for a price quote for a specified date range for a journey from an origin to a destination, the computer terminal further arranged to receive the price quote.

[0182] A good or service is provided, for example an airline ticket, and a price quote for the good or service is provided using the method defined in any aspect of Concept A or the server defined in any aspect of Concept A.

[0183] A web-based price comparison website, wherein an end user can interact with the web-based price comparison website by providing a request for a price for an item or service, the web-based price comparison website A web-based price comparison website is provided that submits a request to a server that quotes prices using a method according to any aspect of Concept A.

[0184] B. How to infer which fare classes are available for travel A method for predicting which fare classes are available for travel from an origin to a destination on a particular date. The method includes: (i) specifying a request for a fare price for travel from an origin to a destination on a particular date; (ii) inferring, estimating, or predicting quoted freight prices from an incomplete historical price dataset by analyzing patterns in the dataset; (iii) using the results of step (ii) to calculate a quote for the requested freight price; (iv) sending a request to a distribution system for fare prices for travel from an origin to a destination on a particular date; (v) receiving from the distribution system a distribution system fare price for travel from the origin to the destination on a particular date; (vi) comparing the calculated quote for the requested fare price from step (iii) with the distribution system fare prices received in step (v) to infer which fare classes are available for travel from the origin to the destination on the particular date; and (vii) outputting the inferred fare class availability for travel from the origin to the destination on the particular date.

[0185] The foregoing may further include any of the following, either alone or in combination: In the method, step (ii) comprises Obtaining historical fare quotes from a computer data store; Grouping past fare quotes by category and Deriving statistical data for each group; storing, on the computer, a classifier including the derived statistics for each group; and identifying, using the stored classifier, a group to which the requested fare price corresponds. In the method, step (iii) includes calculating a set of quotes for the requested freight prices over the specified date range using statistical data obtained from the stored classifiers corresponding to the identified groups. · The distribution system is a global distribution system. An inferred fare class price accompanies each inferred fare class availability. Includes a step to send the output to a server. The server is an airline server. · Step (ii) involves analysing patterns in the data set using the rules. · Step (ii) involves a naive Bayes classifier machine learning approach to generate a probabilistic model of fares, which is then used to predict unobserved fares. · Train a classifier using the observed prices and their corresponding feature sets. · The characteristics relate to the request and include one or more of: day of week departure, length of stay, stay on Saturday, airline, time to travel, route, month. The classifier then predicts the price of an unobserved fare given a set of features and the most likely price with those features. Derive features by training multiple models with different features, This may be done by comparing the predictive accuracy of different models. · Step (ii) includes building a statistical model from historical prices, identifying missing quote candidates, and pricing the quote candidates based on the statistical model. A price estimate is generated for each candidate quote by extracting categorical feature values ​​from the quote, retrieving a trained classifier from a database for the extracted categories, extracting all feature values ​​from the candidate quote, and classifying the candidate quote by computing a Bayes posterior probability for each price range stored in the classifier, selecting the price range class with the highest Bayes posterior probability, and assigning the price class to the candidate quote. · Inputs to the statistical model include a list of pathways, a classifier categorization scheme, past estimates, and a set of supported features with weights. Filter past quotes by age. · Input to the statistical model includes reverse path equivalents. The method includes the step of including the cached freight prices in a set of freight price quotes. The trip is a one-way trip. The trip is a round trip. -Fare is airfare. · The request includes a flexible search request. The estimation process is parameterized by one or more of the following: the minimum Bayes posterior probability required to accept the classification result, the maximum number of path operators involved in generating the candidates, or a random variation added to the Bayes posterior probability to avoid ties. The method is performed on a server where a quote service for fare availability and prices is provided.

[0186] Further, a server configured to infer which fare classes are available for travel from an origin to a destination on a particular date, the server comprising: (i) prescribing requirements for fare prices for travel from an origin to a destination on specified dates; (ii) inferring, estimating, or predicting quoted freight prices from an incomplete historical pricing data set by analyzing patterns in the data set; (iii) calculating an estimate for the requested freight price using the results of the quoted freight rates under (ii); and (iv) sending a request to a distribution system for fare prices for travel from an origin to a destination on a particular date; (v) receiving from the distribution system the distribution system's fare price for travel from the origin to the destination on a particular date; (vi) comparing the calculated estimates for the requested fare prices obtained from (iii) with the distribution system fare prices received in (v) to infer which fare classes will be available for travel from the origin to the destination on a particular date; (vii) outputting the inferred fare class availability for travel from the origin to the destination on a particular date.

[0187] In the server, for (ii), the server (a) obtaining historical fare quotes from a computer data store; (b) grouping past freight quotes by category; and (c) deriving statistical data for each group; and (d) storing, for each group, a classifier including the derived statistics; (e) using the stored classifier to identify the group to which the requested freight price corresponds.

[0188] At the server, for (iii), the server is configured to calculate a set of quotes for the requested freight prices over a specified date range using statistical data obtained from the stored classifiers corresponding to the identified groups.

[0189] The server is further configured to execute a method including any of the method limitations set forth above.

[0190] Further, a computer program product embodied on a non-transitory storage medium, the computer program product, when executed on a computer, is arranged to infer which fare classes are available for travel from an origin to a destination on a particular date, the computer program product, when executed on the computer, (i) prescribing requirements for fare prices for travel from an origin to a destination on specified dates; (ii) inferring, estimating, or predicting quoted freight prices from an incomplete historical pricing data set by analyzing patterns in the data set; (iii) calculating an estimate for the requested freight price using the results of the quoted freight rates under (ii); and (iv) sending a request to a distribution system for fare prices for travel from an origin to a destination on a particular date; (v) receiving from the distribution system the distribution system's fare price for travel from the origin to the destination on a particular date; (vi) comparing the calculated estimates for the requested fare prices obtained from (iii) with the distribution system fare prices received in (v) to infer which fare classes will be available for travel from the origin to the destination on a particular date; (vii) outputting the inferred fare class availability for travel from the origin to the destination for a particular date.

[0191] The computer program product is arranged to perform a method according to any of the preceding method limitations.

[0192] There is provided a computer coupled to a server of any of the server concepts of Concept B above, the computer configured to send a request specifying a journey from an origin to a destination on a particular date, the computer further configured to receive the output inferred fare class availability for the journey from the origin to the destination on the particular date.

[0193] C. How to estimate freight prices 1. A method for estimating a freight price, the method comprising: (i) obtaining historical fare quotes from a computer data store; (ii) grouping the past freight quotes by category; (iii) deriving statistical data for each group; (iv) storing, on the computer, a classifier including the derived statistics for each group; (v) receiving a request for a fare price for travel from an origin to a destination on a particular date; (vi) identifying a group to which the requested fare price corresponds using the stored classifier; (vii) calculating a quote for the requested freight price using statistical data obtained from the stored classifiers corresponding to the identified groups; and (viii) providing the requested freight price quote.

[0194] D. How to estimate freight prices 1. A method for estimating a freight price, the method comprising: (i) receiving a request for a fare price for travel from an origin to a destination on a particular date; (ii) inferring, estimating, or predicting quoted freight rates from an incomplete historical price data set by analyzing patterns in the data set; (iii) using the results of step (ii) to calculate a quote for the requested freight price; (iv) providing the requested freight price quote.

[0195] E. How to Provide a Set of Round-Trip Fare Price Quotes 1. A method for providing a set of round trip fare price quotes, the method comprising: (i) obtaining historical fare quotes from a computer data store; (ii) grouping the past freight quotes by category; (iii) deriving statistical data for each group; (iv) storing, on the computer, a classifier including the derived statistics for each group; (v) receiving a request for round trip fare prices for a specified departure date and a specified return date for travel from an origin to a destination; (vi) identifying a group to which the requested round trip fare price corresponds using the stored classifier; and (vii) calculating a set of quotes for the requested round trip fare prices for the specified departure and return dates using the statistical data obtained from the stored classifiers corresponding to the identified groups; (viii) providing a set of freight price quotes.

[0196] F. Method for Providing the Best Set of Round Trip Fare Price Quotes 1. A method for providing a set of best round trip fare price quotes for a specified departure date range and a specified return date range, the method comprising: (i) obtaining historical fare quotes from a computer data store; (ii) grouping the past freight quotes by category; (iii) deriving statistical data for each group; (iv) storing, on the computer, a classifier including the derived statistics for each group; (v) receiving a request for round trip fare prices for a specified departure date range and a specified return date range for travel from an origin to a destination; (vi) identifying a group to which the requested round trip fare price corresponds using the stored classifier; and (vii) calculating a set of quotes for requested round trip fare prices for pairs of departure and return dates using the statistical data obtained from the stored classifiers corresponding to the identified groups; (viii) selecting the best fare price quote for each pair of departure and return dates; (ix) providing a set of best round trip fare price quotes for each pair of departure and return dates.

[0197] G. How to Provide a Set of Round-Trip Fare Price Quotes 1. A method for providing a set of round trip fare price quotes, the method comprising: (i) receiving a request for round trip fare prices for a specified departure date and a specified return date for travel from an origin to a destination; (ii) inferring, estimating, or predicting quoted freight rates from an incomplete historical price data set by analyzing patterns in the data set; (iii) using the results of step (ii) to calculate a set of quotes for requested round trip fare prices for the specified departure and return dates; (iv) providing a set of freight price quotes.

[0198] H. How to Provide the Best Set of Round-Trip Fare Price Quotes 1. A method for providing a set of best round trip fare price quotes for a specified departure date range and a specified return date range, the method comprising: (i) receiving a request for round trip fare prices for a specified departure date range and a specified return date range for travel from an origin to a destination; (ii) inferring, estimating, or predicting quoted freight rates from an incomplete historical price data set by analyzing patterns in the data set; (iii) using the results of step (ii) to calculate a set of quotes for requested round trip fare prices for pairs of departure and return dates; (iv) providing a set of best round trip fare price quotes for each pair of departure and return dates.

[0199] I. How to guess which fare classes are available. 1. A method for inferring which offered fare classes are available, the method comprising: (i) receiving, by a computer server, a request for goods or services along with parameters defining those goods or services along with a price for those goods or services, e.g., airfare; (ii) at any time relative to step (i), configuring one or more processors to determine an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) configuring one or more processors to calculate estimates for requested prices for goods or services that satisfy the parameters; (iv) sending a request to a distribution system for a freight price; (v) receiving a distribution system freight price from the distribution system; (vi) comparing the calculated estimate for the requested price from step (iii) with the distribution system fare prices received in step (v) to estimate fare class availability; (vii) providing the inferred fare class availability to a computing device.

[0200] The foregoing may further include any of the following, either alone or in combination: Parameters defining those goods or services include one or more of the following: action type, e.g. airfare, hotel booking, train fare; date range; destination; origin; desired weather conditions; star rating; keywords; any other user-defined preferences. · Determining price quotes by inferring, deriving or predicting price quotes. Step (ii) a) obtaining historical price quotes from a computer data store; b) grouping past price quotes by category; and c) deriving statistical data for each group; d) storing, on the computer, a classifier including the derived statistics for each group; e) using the stored classifier to identify the group to which the requested price corresponds. · Step (iii) involves calculating a set of quotes for the required prices over the specified date range using the statistical data obtained from the stored classifiers corresponding to the identified groups. · The distribution system is a global distribution system. An inferred fare class price accompanies each inferred fare class availability. · Step (vii) involves sending the inferred fare class availability to a server. The server is an airline server. · Step (ii) involves analysing patterns in the data set using the rules. · Step (ii) involves a naive Bayes classifier machine learning approach to generate a probabilistic model of prices and use that model to predict unobserved prices. · Train a classifier using the observed prices and their corresponding feature sets. · The characteristics relate to the request and include one or more of: day of week departure, length of stay, stay on Saturday, airline, time to travel, route, month. The classifier then predicts the price of an unobserved price, given a set of features and the most likely price to have those features. Deriving features may be done by training multiple models with different features and comparing the predictive accuracy of the different models. · Step (ii) includes building a statistical model from historical prices, identifying missing quote candidates, and pricing the quote candidates based on the statistical model. A price estimate is generated for each candidate quote by extracting categorical feature values ​​from the quote, retrieving a trained classifier from a database for the extracted categories, extracting all feature values ​​from the candidate quote, and classifying the candidate quote by computing a Bayes posterior probability for each price range stored in the classifier, selecting the price range class with the highest Bayes posterior probability, and assigning the price class to the candidate quote. · Inputs to the statistical model include a list of pathways, a classifier categorization scheme, past estimates, and a set of supported features with weights. Filter past quotes by age. · Input to the statistical model includes reverse path equivalents. The method includes the step of including the cached fare prices in a set of price quotes. The prices are for trips that are one-way trips. ·Prices are for trips which are round trips. The price includes airfare. -Price includes rail freight. The price includes car rental. The price includes hotel rates. · The request includes a flexible search request. The estimation process is parameterized by one or more of the following: the minimum Bayes posterior probability required to accept the classification result, the maximum number of path operators involved in generating the candidates, or a random variation added to the Bayes posterior probability to avoid ties. The method is performed on a server where a quote service for fare availability and prices is provided.

[0201] Further provided is a server configured to perform any of the methods described above.

[0202] Further, a server configured to infer which fare classes are available, the server comprising: (i) receiving a request for goods or services along with parameters defining those goods or services, such as a price for those goods or services, e.g., airfare; (ii) at any time relative to (i) above, configuring one or more processors to determine an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) configuring one or more processors to calculate estimates for requested prices for goods or services that meet the parameters; and (iv) sending the request to a distribution system for a freight price; (v) receiving the distribution system's freight price from the distribution system; and (vi) comparing the calculated estimates for the requested prices obtained from (iii) with the distribution system fare prices received in (v) to estimate fare class availability; (vii) providing the inferred fare class availability to a computing device.

[0203] The foregoing may further include any of the following, either alone or in combination: Parameters defining those goods or services include one or more of the following: action type, e.g. airfare, hotel booking, train fare; date range; destination; origin; desired weather conditions; star rating; keywords; any other user-defined preferences. · Determining price quotes by inferring, deriving or predicting price quotes. For (ii), the server: Obtaining historical price quotes from a computer data store; and Group past price quotes by category and Derive statistical data for each group; storing, for each group, a classifier including the derived statistics; using the stored classifier to identify the group to which the requested price corresponds. For (iii), the server is configured to calculate a set of quotes for the requested price over a specified date range using statistical data obtained from the stored classifiers corresponding to the identified groups. The server is further arranged to perform a method according to any of the method aspects described above.

[0204] A computer program product embodied on a non-transitory storage medium, the computer program product being arranged, when executed on a computer, to infer which fare classes are available, the computer program product being arranged, when executed on the computer, to: (i) The price for goods or services, for example, airfare together with the price for those goods or services receiving a request for parameters defining a service; (ii) at any time relative to step (i), configuring one or more processors to determine an estimated price from an incomplete historical price data set by analyzing patterns in the data set; (iii) configuring one or more processors to calculate estimates for requested prices for goods or services that meet the parameters; and (iv) sending the request to a distribution system for a freight price; (v) receiving the distribution system's freight price from the distribution system; and (vi) comparing the calculated estimate for the requested price from step (iii) with the distribution system fare prices received in step (v) to estimate fare class availability; (vii) providing the inferred fare class availability to a computing device.

[0205] The computer program product may be arranged to perform a method according to any of the method aspects described above.

[0206] There is provided a computer terminal connected to the server of any of the server aspects described above, the computer terminal arranged to send a request specifying fare class availability for goods or services, e.g. airfare, together with a request for parameters defining those goods or services, the computer further arranged to receive the output inferred fare class availability.

[0207] A good or service is provided, for example an airline ticket, for which fare class availability is provided using a method as defined in any of the previously described method aspects or a server as in any of the previously described server aspects.

[0208] A web-based price comparison website is provided where an end user can interact with the web-based price comparison website by providing a request for fare class availability, the web-based price comparison website providing the request to a server which estimates fare class availability using a method according to any of the aspects of the method described above.

[0209] J. How to estimate fare class availability for a particular flight 1. A method for predicting fare class availability for a particular flight, comprising: (a) receiving observable, live, bookable prices for its flights, where that pricing information is not accompanied by complete fare class information for those flights; and (b) comparing observable, live, bookable prices to prices calculated from stored data, e.g., sets of valid itinerary data, fare data, e.g., FROP data, and tax / surcharge data; (c) A method is provided for determining which fare class was used in the observable, live, bookable price by determining which fare class was used in the calculated price that most closely matches the observable, live, bookable price.

[0210] The foregoing may further include any of the following, either alone or in combination: - Get live booking prices from cached data ever obtained can be done. Any or all of the applicable rules or restrictions that apply to a particular fare for a particular flight are specified by a fare class, which includes one or more of the following: a class code, e.g., C (Business full fare) or Y (Economy full fare); a fare type, e.g., First Class or Economy; a category, e.g., eligibility, minimum stay, stopovers, or penalties. An inferred fare class price accompanies each inferred fare class availability. · Step (c) involves sending the inferred fare class availability to the server. The server is an airline server. · Step (b) involves using the rules to analyze patterns in the sets of prices calculated from the stored data. Computing the set of prices calculated from the stored data in step (b) includes building a statistical model from the historical prices, identifying missing candidate quotes, and pricing the candidate quotes based on the statistical model. · Input to the statistical model includes reverse path equivalents.

[0211] Further provided is a server configured to perform the method according to any of the preceding claims.

[0212] Further, a computer program product embodied on a non-transitory storage medium, the computer program product being arranged, when executed on a computer, to infer fare class availability for a particular flight, the computer program product being arranged, when executed on the computer, to: (a) receiving observable, live, bookable prices for its flights, where that pricing information is not accompanied by complete fare class information for those flights; and (b) comparing observable, live, bookable prices to prices calculated from stored data, e.g., sets of valid itinerary data, fare data, e.g., FROP data, and tax / surcharge data; (c) determining which fare class was used at an observable, live, bookable price by determining which fare class was used at a calculated price that most closely matches the observable, live, bookable price. A computer program product is provided that is configured to:

[0213] There is provided a computer terminal connected to a server belonging to any of the server aspects described above, the computer terminal arranged to send a request specifying fare class availability for goods or services, e.g. airfare, together with a request for parameters defining those goods or services, the computer further arranged to receive the outputted inferred fare class availability.

[0214] A good or service is provided, for example an airline ticket, for which fare class availability is provided using a method as defined in any of the previously described method aspects or a server as in any of the previously described server aspects.

[0215] A web-based price comparison web site, wherein an end user may interact with the web-based price comparison web site by providing a request for fare class availability, the web-based price comparison web site responding to the request with a fare class availability request using a method according to any of the aspects of the method described above. A web based price comparison website is provided that provides a server with price quotes.

Claims

1. 1. A computer-implemented method for reducing data storage needs and estimating availability of price classes, comprising: (i) receiving, by a computer server, parameters defining a hotel service and a request for a price for the hotel service; (ii) configuring one or more processors to, at any time relative to step (i), determine quoted prices from an incomplete historical price data set in a computer data store by analyzing patterns in the data set, the incomplete historical price data set consuming less data storage capacity than a complete historical price data set; This step is (a) obtaining historical price quotes from said computer data store; (b) grouping the historical price quotes by category; (c) deriving statistical data for each group; (d) storing, on the computer, a classifier for each group that includes the derived statistical data; (e) identifying a group having a stored classifier corresponding to the requested price; (iii) configuring one or more processors to calculate a quote for the requested price for the hotel service satisfying the parameters, the quote including calculating the requested price using statistical data obtained from the stored classifiers corresponding to the identified groups; (iv) sending said request to a distribution system for hotel services prices; (v) receiving distribution system hotel service prices from the distribution system; (vi) comparing the quote for the requested price from step (iii) with the hotel service prices of the distribution system received in step (v) to estimate availability of hotel service price classes; (vii) providing the inferred hotel service price class availability to a computing device.

2. 2. The method of claim 1, wherein the parameters defining the hotel services include one or more of the following: a date range; a destination; an origin; desired weather conditions; a star rating; and keywords.

3. 3. The method of claim 1 or 2, wherein determining the estimated price is performed by inferring, deriving or predicting the estimated price.

4. 4. A method according to claim 1, wherein said step (iii) includes calculating a set of quotes for the requested price over a specified date range using statistical data obtained from the stored classifiers corresponding to the identified groups.

5. The method according to any one of claims 1 to 4, wherein said distribution system is a global distribution system.

6. A method according to any preceding claim, wherein an inferred hotel service class price is attached to the availability of each inferred hotel service price class.

7. The method according to any of claims 1 to 6, wherein said step (vii) comprises sending said inferred hotel service price class availability to a server.

8. A method according to any preceding claim, wherein step (ii) comprises analysing patterns in the data set using rules.

9. 8. A method according to any preceding claim, wherein step (ii) comprises a naive Bayes classifier machine learning approach to generate a probabilistic model of prices, which is used to predict unobserved prices.

10. 10. The method of claim 1, wherein step (ii) comprises: building a statistical model from historical prices; identifying missing quotation candidates; and pricing the quotation candidates based on the statistical model.

11. A method according to any preceding claim, comprising the step of including cached hotel service prices in said set of price quotes.

12. The method according to any of claims 1 to 11, wherein the observed prices and their corresponding feature sets are used to train a classifier.

13. 13. The method of claim 12, wherein the characteristics relate to the request and include one or more of: departure day of the week, number of days of stay, stay on a Saturday, month.

14. 14. The method of claim 13, wherein the classifier then predicts the price of the unobserved price by giving the set of features and giving the most likely price having those features.

15. 13. The method of claim 12, wherein deriving the features is performed by training a plurality of models with different features and comparing the predictive accuracy of the different models.

16. 16. A method according to any preceding claim, comprising the steps of: estimating a price for each candidate quote by extracting categorical feature values ​​from the quote; retrieving a trained classifier from a database for the extracted categories; extracting all feature values ​​from the candidate quotes; and classifying the candidate quotes by calculating a Bayes posterior probability for each price range stored in the classifier; selecting the price range class with the highest Bayes posterior probability; and assigning the price class to the candidate quote.

17. A method according to any preceding claim, wherein inputs to the statistical model include a classifier categorisation scheme, past estimates and a set of supported features with weights.

18. The method of any preceding claim, wherein the request comprises a flexible search request.

19. 1. A system including a server, a storage computer, and a computer data store, the server configured to infer availability of a price class; The server, (i) receiving a request for parameters defining a hotel service and a requested price for the hotel service; (ii) configuring one or more processors to, at any time relative to (i), determine quoted prices from an incomplete historical price data set stored in the computer data store by analyzing statistical patterns in the data set, the incomplete historical price data set consuming less data storage capacity than a complete historical price data set; This configuration is as follows: (a) obtaining historical price quotes from said computer data store; (b) grouping the historical price quotes by category; (c) deriving statistical data for each group; (d) storing, on the storage computer, a classifier for each group that includes the derived statistical data; (e) identifying a group having a stored classifier corresponding to the requested price; (iii) configuring one or more processors to calculate a quote for the requested price for the hotel service satisfying the parameters, including calculating the requested price quote using statistical data obtained from the stored classifiers corresponding to the identified groups; (iv) sending said request to a distribution system for hotel services prices; (v) receiving distribution system hotel service prices from the distribution system; (vi) comparing the quote for the requested price from step (iii) with the hotel service prices of the distribution system received in step (v) to estimate availability of hotel service price classes; (vii) providing the inferred hotel service price class availability to a computing device; The system is configured as follows:

Citation Information

Patent Citations

  • Method and device for receiving order for reservation and recording medium therefor

    JP2001222582A

  • Electronic commerce server and electronic commerce method

    JP2003256701A

  • Method and system for presenting rates for travel services

    JP2008503803A

  • Travel-related prediction system

    US20120239455A1