Order matching method and system of service platform

Abstract

The application provides an order matching method and system of a service platform, acquires user order demand information (including itinerary, service and resource demand) and service providing information (including service provider itinerary, service and resource providing information) in a tourism service platform, then extracts demand characteristics of the user order demand information to obtain a user preference characteristic set, extracts providing characteristics of the service providing information to obtain a service provider capability characteristic set, then performs characteristic correlation processing on both to generate a demand capability correlation matrix, calculates an order service matching degree parameter based on the matrix to represent the degree to which the service providing information satisfies the user order demand information, and finally generates an order matching scheme according to the matching degree parameter and sends it to a user interface for display, thereby improving order matching accuracy and efficiency and meeting user individualized needs.

Description

Technical Field

[0001] This invention relates to the field of digital service technology, and more specifically, to an order matching method and system for a service platform. Background Technology

[0002] In today's rapidly developing tourism service industry, tourism service platforms have become a crucial bridge connecting users and service providers. However, existing tourism service platforms have many shortcomings in order matching. On the one hand, when traditional platforms obtain user order demand information and service provision information, they often simply list the information without systematically classifying it. For example, for user itinerary demand information, they may only record the departure and destination points, ignoring key details such as preferred modes of transportation and length of stay; for service provider itinerary provision information, they only roughly describe the range of itineraries available, without detailing specific time arrangements and itinerary details.

[0003] On the other hand, in terms of feature extraction, existing technologies typically focus only on single-dimensional features, such as considering only users' price preferences for accommodation, while ignoring preferences for itineraries, services, and other aspects, as well as the corresponding capabilities of service providers. This results in an inability to comprehensively and accurately measure the fit between user needs and service offerings during the matching process. Moreover, traditional platforms often use simple algorithms such as weighted summation when calculating the matching degree, failing to delve into the intrinsic relationship between user preference features and service provider capability features. This leads to inaccurate matching results, failing to meet the increasingly diverse and personalized travel needs of users, and also affecting the service efficiency of service providers and the overall operational effectiveness of the platform. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide an order matching method for a service platform, the method comprising: The system obtains user order demand information and service provision information from a tourism service platform. The user order demand information includes user itinerary demand information, user service demand information, and user resource demand information. The service provision information includes service provider itinerary provision information, service provider service provision information, and service provider resource provision information. The user order demand information is processed by demand feature extraction to obtain a user preference feature set, which includes trip preference features, service preference features and resource preference features; The service provision information is processed by feature extraction to obtain a service provider capability feature set, which includes trip provision capability features, service provision capability features and resource provision capability features; The user preference feature set and the service provider capability feature set are subjected to feature association processing to generate a demand capability association matrix. Each element in the demand capability association matrix is ​​used to represent the degree of association between a user preference feature in the user preference feature set and a service provider capability feature in the service provider capability feature set. Based on the demand capability association matrix, the matching degree of the user order demand information and the service provision information is calculated to obtain the order service matching degree parameter. The order service matching degree parameter is used to indicate the degree to which the service provision information meets the user order demand information. An order matching scheme is generated based on the order service matching parameters, and the order matching scheme is sent to the user interface of the tourism service platform for display.

[0005] In another aspect, embodiments of the present invention also provide an order matching system for a service platform, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0006] Based on the above, this embodiment of the invention comprehensively acquires and categorizes user order demand information and service provision information from a tourism service platform, covering user itineraries, service and resource needs, as well as the corresponding service provider provision information. It extracts demand features from user order demand information to obtain a set of user preference features including itinerary, service, and resource preferences. Simultaneously, it extracts provision features from service provision information to obtain a set of service provider capability features. This achieves in-depth mining of multi-dimensional features of users and service providers. Through feature association processing, a demand-capability association matrix is ​​generated, clearly presenting the degree of correlation between user preference features and service provider capability features. It accurately grasps the matching relationship between the two at a micro level. Based on this demand-capability association matrix, a matching degree calculation is performed. The resulting order service matching degree parameter scientifically and accurately represents the degree to which the service provision information meets the user's order demand information. Finally, an order matching scheme is generated based on the matching degree parameter and displayed on the user interface, greatly improving the accuracy and efficiency of order matching, effectively meeting users' personalized needs, and enhancing the service effectiveness of service providers and the overall operational quality of the platform. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of the execution flow of the order matching method of the service platform provided in this embodiment of the invention.

[0008] Figure 2 This is a schematic diagram of exemplary hardware and software components of the order matching system of the service platform provided in this embodiment of the invention. Detailed Implementation

[0009] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating an order matching method for a service platform provided in one embodiment of the present invention. The order matching method for this service platform will be described in detail below.

[0010] Step S110: Obtain user order demand information and service provision information from the tourism service platform. The user order demand information includes user itinerary demand information, user service demand information, and user resource demand information. The service provision information includes service provider itinerary provision information, service provider service provision information, and service provider resource provision information.

[0011] In this embodiment, user order request information and service provision information can be extracted from the database and related storage modules of the tourism service platform. User order request information covers various tourism service requests submitted by the user on the platform. User itinerary request information involves travel arrangements, such as departure point, destination, transit points, and itinerary theme for a specific time period. User service request information includes requirements for services during the tour, such as whether a tour guide or translation service is needed and the service standards. User resource request information refers to requirements for tourism resources, such as hotel type, number of rooms, and vehicle type and quantity.

[0012] Service provision information refers to the tourism services offered by service providers on the platform. Service provider itinerary information includes the available travel time periods, departure points, transit points, destinations, and itinerary themes. Service provider service information covers the types of services offered, such as tour guides, translators, and restaurant reservations, as well as service standards and delivery methods, such as full-time accompaniment or phased services. Service provider resource information includes the types of resources available, such as hotels, vehicles, and attraction tickets, as well as the quantity and quality of the resources.

[0013] In step S110, data collection and data authorization strictly comply with relevant laws and regulations. Before obtaining user order request information, the travel service platform will clearly inform users of the scope, purpose, use, and storage period of data collection through user registration agreements, privacy policies, and other documents. This includes the collection methods and usage scenarios for personal data such as user itinerary request information, service request information, and resource request information. When users submit order requests on the platform, they must actively check the authorization option to agree to data collection and use. This authorization is recorded and archived in written form (electronic agreement) to ensure that users are clearly aware of and voluntarily authorize the platform to process their personal data. Without the user's explicit authorization, the platform will not collect any user information without authorization. Regarding the information provided for services, the platform requires service providers to submit valid business qualification certificates during registration and sign a data provision authorization agreement, declaring that the service information and related data provided are true and legal, do not infringe on the rights and interests of third parties, and authorizing the platform to process and display this information as necessary. During the data collection process, the scope of collection is strictly limited, collecting only data directly related to matching tourism service orders, and excluding sensitive information unrelated to the service. Core sensitive data such as user ID numbers and contact information collected are anonymized, except for necessary steps such as identity verification and order confirmation, removing identifiers that can directly identify the user. The platform has established a data access permission management mechanism, allowing only authorized personnel to access relevant data while performing their duties, and all access activities are fully recorded and traceable. Furthermore, the platform regularly conducts compliance audits of the data collection and authorization processes, inviting third-party organizations to assess the legality of data processing activities to ensure compliance with laws and regulations. If laws and regulations are updated, the platform will promptly adjust its data collection and authorization mechanisms to maintain compliance.

[0014] Step S120: Extract user order demand information to obtain a user preference feature set, which includes trip preference features, service preference features and resource preference features.

[0015] After obtaining user order request information, key features are extracted from the text, data, and other content. These key features are combined to form a set of user preference features, where trip preference features reflect users' preferences in trip planning, service preference features reflect users' preferences in services, and resource preference features show users' preferences in resources.

[0016] Step S121: Perform trip feature parsing processing on the user's trip demand information, extract the trip time feature, trip location feature and trip type feature from the user's trip demand information, and combine the trip time feature, the trip location feature and the trip type feature to form trip preference features.

[0017] The system analyzes user travel needs information to extract travel feature characteristics. This analysis is conducted in depth to extract features related to time, location, and type. These features specifically reflect user preferences in terms of travel, and are combined to form travel preference features.

[0018] Step S1211: Perform text segmentation on the user's travel request information, decomposing the user's travel request information into multiple travel text words.

[0019] User travel needs are mostly presented in text form, such as "I plan to depart from location X, pass through location Y, and arrive at location Z during a specific time period for a historical and cultural tour. Along the way, I will participate in visits to historical sites and folk experience activities. The travel distance is within a certain range."

[0020] The above text was segmented using a dictionary-based word segmentation method combined with a tourism-related professional dictionary. The continuous text was divided into multiple itinerary text words such as "plan", "at", "specific time period", "from", "location X", "departure", "passing through", "location Y", "go to", "location Z", "conduct", "a", "have", "history and culture", "theme", "of", "tourism", "on the way", "will", "participate", "visit historical sites", "and", "folk experience", "activity", "itinerary", "distance", "in a certain range" and "a certain range".

[0021] Step S1212: Perform part-of-speech tagging on the multiple trip text words to identify words indicating time, words indicating location, and words indicating type.

[0022] After obtaining multiple trip text words, a statistical machine learning model, such as a Hidden Markov Model (HMM), is used to tag each word with its part of speech. This machine learning model is trained on a large amount of labeled corpus and can determine the part of speech of a word based on its contextual information.

[0023] In the above examples, after part-of-speech tagging, "specific time period" was identified as a word indicating time; "location X", "location Y", and "location Z" were identified as words indicating location; and "historical and cultural theme", "visiting historical sites", "folk custom experience", and "a certain range" were identified as words indicating type.

[0024] Step S1213: Extract trip time features from time-related words, the trip time features including trip start time, trip end time, and trip duration.

[0025] Trip time features are extracted from the identified time-related words. For the term "specific time period," the specific time point and duration are further determined by combining it with other relevant descriptions in the user's order request information. For example, if the text also mentions "the trip is expected to start at the beginning of the month and end in the middle of the month," then the trip start time is a specific date at the beginning of the month, the trip end time is a specific date in the middle of the month, and the trip duration is the number of days between the start and end times.

[0026] Step S1214: Extract trip location features from the vocabulary representing locations, the trip location features including the trip departure point, the trip route points and the trip destination point.

[0027] Extract itinerary location features from words representing locations. "Location X" is explicitly the departure point of the itinerary, "Location Y" is a location along the route, and "Location Z" is the destination. If the text provides more detailed descriptions of these locations, such as "Location X is a central square in a city," "Location Y is the entrance to an ancient town," or "Location Z is a mountain resort," then this detailed information is included in the corresponding itinerary location features.

[0028] Step S1215: Extract trip type features from the vocabulary representing types, the trip type features including trip theme type, trip activity type and trip distance type.

[0029] Extract itinerary type features from the vocabulary indicating type. "Historical and cultural theme" is identified as the itinerary theme type; "visiting historical sites" and "experiencing folk customs" belong to the itinerary activity type; "a certain range" combined with the specific context is identified as the itinerary distance type, such as medium distance range.

[0030] Step S1216: Combine the trip time feature, the trip location feature, and the trip type feature according to a preset feature combination rule to form a trip preference feature that includes time dimension, location dimension, and type dimension.

[0031] According to the preset feature combination rules, the trip time feature, trip location feature, and trip type feature are combined. For example, first, the trip start time, trip end time, and trip duration are arranged by time dimension; then, the trip departure point, trip route points, and trip destination are arranged by location dimension; finally, the trip theme type, trip activity type, and trip distance type are arranged by type dimension. The combination forms a multi-dimensional trip preference feature.

[0032] Step S1217: Perform feature verification processing on the combined trip preference features to ensure that the trip preference features can accurately reflect the trip preferences in the user's trip demand information.

[0033] The combined trip preference features undergo feature verification processing. By comparing the extracted trip preference features with the original user trip demand information, it is checked for any missing or incorrect feature items. If a feature item is found to be inconsistent with the original information, such as an error in calculating the trip duration, the feature item is re-extracted and combined until the trip preference features accurately reflect the trip preferences in the user trip demand information.

[0034] Step S122: Perform service feature parsing processing on the user service demand information, extract service type features, service standard features and service method features from the user service demand information, and combine the service type features, the service standard features and the service method features to form service preference features.

[0035] The user's service request information is analyzed and processed to identify service characteristics. A user's service request information might be something like, "I want to receive professional tour guide services during my trip, and I require the guide to have extensive historical and cultural knowledge, and the service method is full-time accompaniment."

[0036] First, the text was segmented to obtain words such as "hope", "in", "tourism", "in the process", "obtain", "professional", "guide", "service", "requirement", "guide", "possess", "rich", "of", "history and culture", "knowledge", "service", "method", "for", "whole process", and "accompany".

[0037] Next, part-of-speech tagging was performed to identify the term "tour guide service" which indicates the service type, "professional" and "possesses rich historical and cultural knowledge" which indicate the service standard, and "accompanied throughout the entire process" which indicates the service method.

[0038] The service type characteristics extracted from these words are tour guide service, the service standard characteristics are professionalism and rich historical and cultural knowledge, and the service method characteristics are full-time accompaniment.

[0039] According to the preset combination rules, service type features, service standard features and service method features are combined to form service preference features, which include information in three dimensions: service type, service standard and service method.

[0040] Step S123: Perform resource feature parsing processing on the user resource demand information, extract resource type features, resource quantity features and resource quality features from the user resource demand information, and combine the resource type features, the resource quantity features and the resource quality features to form resource preference features.

[0041] The user's resource requirement information is processed through resource feature analysis. The user's resource requirement information might be "Need to book two rooms in a three-star hotel and rent a comfortable car".

[0042] The text was segmented to obtain words such as "need", "book", "three-star", "hotel", "two rooms", "rent", "comfort type", "car", and "a car".

[0043] After part-of-speech tagging, the words "hotel" and "car" indicating resource type were identified, "two rooms" and "one car" indicating resource quantity were identified, and "three-star" and "comfort type" indicating resource quality were identified.

[0044] The resource type features are hotels and cars, the resource quantity features are two hotels and one car, and the resource quality features are three-star hotels and comfort cars.

[0045] These features are combined according to the combination rules to form resource preference features, which include information in three dimensions: resource type, resource quantity, and resource quality.

[0046] Step S124: Perform feature standardization and integration processing on the trip preference features, service preference features and resource preference features to form a set of user preference features that include trip dimension, service dimension and resource dimension.

[0047] The trip preference features, service preference features, and resource preference features are standardized and integrated. The Min-Max standardization method is used to map the numerical range of each feature to the same interval, for example, transforming all feature values ​​to the range of 0-1.

[0048] For time features in travel preference features, if the original time is in days and ranges from 1 to 10 days, it is converted into a value between 0 and 1 using a standardization formula; for non-numerical features such as location features and type features, one-hot encoding is used to process them and convert them into numerical features.

[0049] The standardized trip preference features, service preference features, and resource preference features are integrated in the order of trip dimension, service dimension, and resource dimension to form a user preference feature set. This user preference feature set is a multi-dimensional feature combination that comprehensively reflects the user's preferences in terms of trip, service, and resources.

[0050] Step S130: Perform feature extraction processing on the service provision information to obtain a service provider capability feature set, which includes trip provision capability features, service provision capability features, and resource provision capability features.

[0051] After obtaining service provision information, feature extraction processing is performed. Using appropriate algorithms and rules, key features are extracted from the service provision information. These features are combined to form a set of service provider capability features. Among them, the trip provision capability feature reflects the service provider's ability to arrange trips, the service provision capability feature reflects the service provider's ability to provide services, and the resource provision capability feature demonstrates the service provider's ability to provide resources.

[0052] Step S131: Perform trip provision capability parsing processing on the service provider's trip provision information, extract the trip provision time feature, trip provision location feature and trip provision type feature from the service provider's trip provision information, and combine the trip provision time feature, the trip provision location feature and the trip provision type feature to form trip provision capability feature.

[0053] Perform itinerary offering capability analysis and processing on the service provider's itinerary offering information. The service provider's itinerary offering information may be "able to arrange an itinerary in the first half of each month, starting from location A, passing through location B, and going to location C, offering an itinerary with a historical and cultural theme".

[0054] Perform word segmentation on this text to obtain words such as "able to", "in", "each month", "the", "first half", "arrange", "itinerary", "able to", "from", "location A", "depart", "pass through", "location B", "go to", "location C", "offer", "historical and cultural", "theme", "the", "itinerary", etc.

[0055] After performing词性标注 (it should be "pos tagging" in English), identify the words indicating the offering time "the first half of each month", the words indicating the offering locations "location A", "location B", "location C", and the words indicating the offering type "historical and cultural theme".

[0056] Extract the itinerary offering time feature as the first half of each month from these words, the itinerary offering location features include the departure location A, the passing-through location B, and the destination location C, and the itinerary offering type feature is the historical and cultural theme.

[0057] According to the preset feature combination rules, combine these features to form the itinerary offering capability feature, which contains information in three dimensions: offering time, offering location, and offering type.

[0058] For example, step S1311: Perform word segmentation on the service provider's itinerary offering information, and decompose the service provider's itinerary offering information into multiple itinerary offering text words.

[0059] The service provider's itinerary offering information is in text form, such as "able to arrange an itinerary starting from location D, passing through location E, and going to location F with a natural landscape theme in the first two months of each quarter".

[0060] Adopt a dictionary-based word segmentation method and combine it with a professional dictionary in the tourism field to perform word segmentation on this text, obtaining multiple itinerary offering text words such as "able to", "in", "each quarter", "the", "first two months", "arrange", "from", "location D", "depart", "pass through", "location E", "go to", "location F", "the", "natural landscape", "theme", "itinerary", etc.

[0061] Step S1312: Perform pos tagging processing on the multiple itinerary offering text words to identify the words indicating the offering time, the words indicating the offering locations, and the words indicating the offering type.

[0062] It should be noted that "词性标注" in Chinese is usually translated as "pos tagging" in English in the context of natural language processing. If there are more specific requirements or corrections for this translation, please let me know.Hidden Markov Models (HMMs) were used to perform part-of-speech tagging on the vocabulary of multiple trip provision texts. In the example above, "the first two months of each quarter" was identified as a word indicating the provision time, "location D", "location E", and "location F" were identified as words indicating the provision location, and "natural landscape theme" was identified as a word indicating the provision type.

[0063] Step S1313: Extract trip provision time features from the vocabulary representing the provision time, the trip provision time features including trip provision start time, trip provision end time and trip provision duration.

[0064] The time characteristics of the trip offering are extracted from the term "the first two months of each quarter". The trip offering starts on the first day of the first month of each quarter, ends on the last day of the second month of each quarter, and lasts for the total number of days in these two months.

[0065] Step S1314: Extract trip location features from the vocabulary representing the location provided, the trip location features including the trip departure point, the trip transit points and the trip destination point.

[0066] Extract itinerary location features from the vocabulary indicating the locations provided. "Location D" provides the departure point of the itinerary, "Location E" provides the transit points, and "Location F" provides the destination point. If there is a more detailed description, such as "Location D is a train station square," then that detailed information is included in the itinerary departure point feature.

[0067] Step S1315: Extract trip provision type features from the vocabulary representing the provision type, the trip provision type features including trip provision theme type, trip provision activity type and trip provision distance type.

[0068] Extracting the itinerary provision type characteristics from the term "natural landscape theme," the itinerary provision theme type is determined to be "natural landscape theme." Combining this with other descriptions in the service provider's itinerary provision information, if it mentions "hiking and picnic activities can be arranged along the way, within a certain distance," then the itinerary activity type is hiking and picnic, and the itinerary distance type is "within that certain distance."

[0069] Step S1316: Combine the trip provision time feature, the trip provision location feature, and the trip provision type feature according to a preset feature combination rule to form a trip provision capability feature that includes a provision time dimension, a provision location dimension, and a provision type dimension.

[0070] According to the preset feature combination rules, first arrange the travel start time, travel end time, and travel duration in the travel time feature provided by the itinerary, then arrange the departure location, passing locations, and destination location in the travel location feature provided by the itinerary, and finally arrange the travel theme type, travel activity type, and travel distance type in the travel type feature provided by the itinerary, and combine them to form the travel providing ability feature.

[0071] Step S132: Perform service providing ability analysis and processing on the service provider's service providing information, extract the service providing type feature, service providing standard feature, and service providing method feature in the service provider's service providing information, and combine the service providing type feature, the service providing standard feature, and the service providing method feature to form the service providing ability feature.

[0072] Perform service providing ability analysis and processing on the service provider's service providing information. The service provider's service providing information may be "able to provide professional translation services, the translators need to be proficient in multiple languages, and the service method is available on call".

[0073] After word segmentation of this text, words such as "able to", "provide", "professional", "translation", "service", "translator", "need to", "be proficient in", "multiple", "languages", "service", "method", "is", "available on call" are obtained.

[0074] After词性标注, identify the word "translation service" indicating the service providing type, the words "professional" and "proficient in multiple languages" indicating the service providing standard, and the word "available on call" indicating the service providing method.

[0075] Extract the service providing type feature as translation service, the service providing standard feature as professional and proficient in multiple languages, the service providing method feature as available on call, and combine them to form the service providing ability feature.

[0076] Step S133: Perform resource providing ability analysis and processing on the service provider's resource providing information, extract the resource providing type feature, resource providing quantity feature, and resource providing quality feature in the service provider's resource providing information, and combine the resource providing type feature, the resource providing quantity feature, and the resource providing quality feature to form the resource providing ability feature.

[0077] Perform resource providing ability analysis and processing on the service provider's resource providing information. The service provider's resource providing information may be "able to provide three four-star hotels and two luxury buses".

[0078] After word segmentation, words such as "able to", "provide", "four-star", "hotel", "three", "luxury", "bus", "two" are obtained. It should be noted that the term "词性标注" in the original text seems to be a Chinese term that might need to be adjusted to the appropriate English expression related to word tagging in the context. Here, it is left as is for the purpose of following the translation rules strictly. You may want to double-check and correct it if necessary in a more comprehensive context.

[0079] After part-of-speech tagging, words indicating the type of resource provision, such as "hotel" and "bus", words indicating the quantity of resource provision, such as "three rooms" and "two vehicles", and words indicating the quality of resource provision, such as "four-star" and "luxury".

[0080] The resource provision type characteristics are hotels and buses; the resource provision quantity characteristics are three hotels and two buses; and the resource provision quality characteristics are four-star hotels and luxury buses. These characteristics are combined to form the resource provision capacity characteristics.

[0081] Step S134: Perform feature standardization and integration processing on the trip provision capability features, the service provision capability features, and the resource provision capability features to form a service provider capability feature set that includes trip provision dimension, service provision dimension, and resource provision dimension.

[0082] The features of trip provision capability, service provision capability, and resource provision capability are standardized and integrated. The same standardization method used for the user preference feature set, namely Min-Max standardization, is employed to map the numerical ranges of each feature to the same interval. For non-numerical features, one-hot encoding is also used to convert them into numerical features.

[0083] The integration process proceeds in the order of trip provision, service provision, and resource provision. The standardized trip provision capability characteristics, service provision capability characteristics, and resource provision capability characteristics are then sequentially arranged and combined to form a service provider capability characteristic set. This set comprehensively reflects the service provider's capabilities in trip provision, service provision, and resource provision. Each dimension's characteristics have been standardized to ensure comparability between different characteristics.

[0084] Step S140: Perform feature association processing on the user preference feature set and the service provider capability feature set to generate a demand capability association matrix. Each element in the demand capability association matrix is ​​used to represent the degree of association between a user preference feature in the user preference feature set and a service provider capability feature in the service provider capability feature set.

[0085] After obtaining the user preference feature set and the service provider capability feature set, feature association processing is performed. By analyzing the inherent relationships between the features in the two sets, the degree of association between each user preference feature and each service provider capability feature is determined, and these degrees of association are represented in matrix form to form a demand-capability association matrix.

[0086] Step S141: Determine the number of user preference features included in the user preference feature set and the number of service provider capability features included in the service provider capability feature set.

[0087] Count the number of user preference features in the user preference feature set. The user preference feature set includes trip preference features, service preference features, and resource preference features. Each preference feature contains multiple sub-features. For example, trip preference features include trip time features, trip location features, and trip type features. Count the total number of all these sub-features, denoted as U.

[0088] Similarly, count the number of service provider capability features in the service provider capability feature set. The service provider capability feature set includes trip provision capability features, service provision capability features, and resource provision capability features. Each capability feature also contains multiple sub-features. Count the total number of all sub-features, denoted as S.

[0089] Step S142: Construct an initial association matrix based on the number of user preference features and the number of service provider capability features. The number of rows in the initial association matrix is ​​the same as the number of user preference features, and the number of columns in the initial association matrix is ​​the same as the number of service provider capability features.

[0090] Based on the number of user preference features U and the number of service provider capability features S, construct an initial association matrix with U rows and S columns. All elements in the initial association matrix are set to 0. The rows of this initial association matrix correspond to each user preference feature in the user preference feature set, and the columns correspond to each service provider capability feature in the service provider capability feature set.

[0091] Step S143: For each user preference feature in the user preference feature set, calculate the correlation value between the user preference feature and each service provider capability feature in the service provider capability feature set.

[0092] For the first user preference feature in the user preference feature set, the correlation degree value is calculated sequentially with each service provider capability feature in the service provider capability feature set. After the calculation is completed, the same operation is performed for the second user preference feature in the user preference feature set, until the correlation degree value between all user preference features and all service provider capability features has been calculated.

[0093] Step S1431: Obtain the feature attribute description of the user preference feature and the feature attribute description of the service provider capability feature.

[0094] Extract the feature attribute descriptions of user preference features. These feature attribute descriptions are specific explanations of user preference features. For example, the attribute description of the travel time feature might be "hopes to travel between the beginning and middle of the month".

[0095] At the same time, extract the characteristic attribute descriptions of the service provider's capabilities, such as the attribute description of the trip provision time characteristic, which may be "trips can be arranged in the first half of each month".

[0096] Step S1432: Compare the feature attribute descriptions of the user preference features and the feature attribute descriptions of the service provider capability features to determine the attribute matching items and attribute difference items between the two.

[0097] Comparing the two feature attribute descriptions above, there is an overlapping time period between "beginning of the month to the middle of the month" and "the first half of each month". This overlapping content is the attribute matching item; while the time period in "beginning of the month to the middle of the month" that exceeds "the first half of each month" and the part in "the first half of each month" that may not be covered by "beginning of the month to the middle of the month" are attribute difference items.

[0098] Step S1433: Calculate the attribute matching rate based on the number of attribute matching items and the number of attribute difference items. The attribute matching rate is the ratio of the number of attribute matching items to the sum of the number of attribute matching items and the number of attribute difference items.

[0099] The number of statistically matched attributes is denoted as M; the number of statistically differing attributes is denoted as N. The attribute matching rate is calculated by dividing M by (M plus N), and the result is the attribute matching rate.

[0100] Step S1434: Based on the attribute matching rate and combined with the preset attribute weight coefficient, calculate the basic correlation value between the user preference feature and the service provider capability feature.

[0101] A preset attribute weight coefficient is set, which is determined based on the importance of the attribute. For example, the time attribute is more important in itinerary planning, so a higher weight coefficient can be set, denoted as W. The basic correlation degree value is the attribute matching rate multiplied by W.

[0102] Step S1435: Adjust the basic correlation degree value by weighting the basic correlation degree value according to the importance coefficient of the user preference feature and the importance coefficient of the service provider capability feature to obtain the final correlation degree value.

[0103] Assign an importance coefficient, denoted as Pu, to the user preference feature. This coefficient reflects the importance of the user preference feature within the overall user preferences; the higher the importance, the larger the coefficient. Assign an importance coefficient, denoted as Ps, to the service provider capability feature. This coefficient reflects the importance of the service provider capability feature within the overall service provider capabilities. The final correlation value is calculated by multiplying the base correlation value by Pu and then by Ps.

[0104] Step S144: Fill the calculated correlation degree value into the corresponding position in the initial correlation matrix to generate a demand capability correlation matrix containing the correlation degree values ​​of all user preference features and service provider capability features.

[0105] The calculated correlation values ​​between each user preference feature and each service provider capability feature are then entered into the corresponding rows and columns of the initial correlation matrix. After all features are entered, the initial correlation matrix is ​​transformed into a demand-capability correlation matrix, which fully represents the correlation between all user preference features and service provider capability features.

[0106] Step S150: Based on the demand capability association matrix, perform matching degree calculation on the user order demand information and the service provision information to obtain the order service matching degree parameter. The order service matching degree parameter is used to indicate the degree to which the service provision information meets the user order demand information.

[0107] By utilizing the correlation values ​​in the demand-capacity correlation matrix and through a series of calculation steps, we obtain the order service matching degree parameter, which represents the degree to which the information provided by the service meets the user's order requirements.

[0108] Step S151: Obtain the preference weight value of each user preference feature in the user preference feature set and the capability weight value of each service provider capability feature in the service provider capability feature set.

[0109] Assign a preference weight value to each user preference feature in the user preference feature set. The preference weight value is determined according to the user's importance to the feature. The more important the feature is to the user, the higher the preference weight value. The sum of the preference weight values ​​of all user preference features is 1.

[0110] Similarly, a capability weight value is assigned to each service provider capability feature in the service provider capability feature set. This capability weight value is determined based on the degree of advantage the service provider has in that capability feature. The more obvious the advantage of the feature, the higher the capability weight value. The sum of the capability weight values ​​of all service provider capability features is 1.

[0111] Step S152: Based on the preference weight value and the capability weight value, the correlation degree value in the demand capability correlation matrix is ​​weighted to obtain the weighted demand capability correlation matrix.

[0112] For each element in the demand-capability correlation matrix, i.e., the correlation degree value, it is multiplied by the preference weight value of the corresponding user preference feature and the capability weight value of the corresponding service provider capability feature to obtain the weighted correlation degree value. After all elements are updated, the weighted demand-capability correlation matrix is ​​formed.

[0113] Step S153: Perform row-wise summation on the weighted demand capability association matrix to obtain the total association degree value corresponding to each user preference feature.

[0114] In the weighted demand capability association matrix, the sum of all element values ​​in each row is the total association value of the user preference feature corresponding to that row.

[0115] Step S154: Perform column-wise summation on the weighted demand capability association matrix to obtain the total association degree value corresponding to each service provider capability feature.

[0116] In the weighted demand capability association matrix, the sum of all element values ​​in each column is the total association value of the service provider capability characteristics corresponding to that column.

[0117] Step S155: Sum the total correlation values ​​corresponding to all user preference features to obtain the total correlation value of user preferences, and sum the total correlation values ​​corresponding to all service provider capability features to obtain the total correlation value of service provider capabilities.

[0118] The total correlation values ​​of all user preference features obtained in step S153 are summed to obtain the total correlation value of user preferences, denoted as T1.

[0119] The total correlation values ​​of all service provider capability features obtained in step S154 are summed to obtain the total correlation value of service provider capabilities, denoted as T2.

[0120] Step S156: Calculate the order service matching degree parameter based on the total correlation degree value of user preferences and the total correlation degree value of service provider capabilities. The order service matching degree parameter is the ratio of the product of the total correlation degree value of user preferences and the total correlation degree value of service provider capabilities to the sum of the two.

[0121] The order service matching degree parameter is calculated by dividing (T1 multiplied by T2) by (T1 plus T2), and the result is the order service matching degree parameter.

[0122] Step S160: Generate an order matching scheme based on the order service matching degree parameters, and send the order matching scheme to the user interface of the tourism service platform for display.

[0123] Based on the order service matching parameter, determine whether to generate an order matching plan and send the generated plan to the user interface of the travel service platform for users to view.

[0124] Step S161: Obtain a preset order matching threshold and compare the order service matching degree parameter with the order matching threshold.

[0125] Retrieve a preset order matching threshold from the configuration file or relevant storage area of ​​the travel service platform. This threshold is the standard for determining whether the service provided information matches the user's order requirements. Compare the calculated order service matching degree parameter with this order matching threshold.

[0126] Step S162: When the order service matching degree parameter is greater than or equal to the order matching threshold, it is determined that the service provision information and the user order demand information are successfully matched, and a preliminary order matching scheme is generated based on the service provision information.

[0127] If the order service matching parameter is greater than or equal to the order matching threshold, it means that the service information provided can meet the user's order requirements. At this time, a preliminary order matching plan is generated based on the service information provided. This preliminary order matching plan includes preliminary information such as itinerary arrangements, service content, and resource configuration, such as the specific time schedule of the itinerary, the list of services provided, and the types and quantities of resources provided.

[0128] Step S163: When the order service matching degree parameter is less than the order matching threshold, it is determined that the service provision information and the user order demand information fail to match, and the process returns to the step of obtaining service provision information, re-obtaining new service provision information and calculating the matching degree.

[0129] If the order service matching degree parameter is less than the order matching threshold, it indicates that the service provided information cannot meet the user's order requirements. At this time, return to step S110, retrieve new content from the service provided information, and then repeat the above steps to extract features, perform association processing and match degree calculation until a matching service provided information is found.

[0130] Step S164: Optimize the preliminary order matching scheme by adjusting the itinerary, service content, and resource allocation in the preliminary order matching scheme based on the user preference feature set and the service provider capability feature set.

[0131] Step S1641: Adjust the itinerary arrangement in the preliminary order matching scheme according to the itinerary preference characteristics and the itinerary provision capability characteristics, and optimize the itinerary time sequence and itinerary location connection.

[0132] Based on user time and location preferences in the trip preference characteristics and the service provider's available time and location arrangements in the trip provision capability characteristics, the order of trip times in the initial order matching scheme is adjusted to ensure that the trip times match user preferences and are within the service provider's available timeframe. Simultaneously, connections between trip locations are optimized to reduce unnecessary travel time and make the trip smoother.

[0133] Step S1642: Adjust the service content in the preliminary order matching scheme according to the service preference characteristics and the service provision capability characteristics, add service items that meet the user's service needs, and reduce service items that do not meet the user's service needs.

[0134] Based on the user's requirements for service type, standard and method in the service preference characteristics, and the service content that the service provider can provide in the service provision capability characteristics, the initial order matching scheme adds service items that the user needs but are not included, and deletes service items that the user does not need, so that the service content is more in line with the user's needs.

[0135] Step S1643: Adjust the resource configuration in the preliminary order matching scheme according to the resource preference characteristics and the resource provision capability characteristics, and optimize the resource allocation quantity and resource usage order.

[0136] Based on user preferences regarding resource type, quantity, and quality, and the resource availability provided by service providers, the resource configuration in the initial order matching scheme is adjusted. For example, the quantity of a certain type of resource required by the user may be increased, or the order in which resources are used may be adjusted to improve resource utilization efficiency.

[0137] Step S1644: Conduct a comprehensive evaluation of the adjusted itinerary, service content, and resource allocation, and calculate the overall satisfaction parameter after the plan adjustment.

[0138] A comprehensive evaluation index system was established, which includes indicators such as itinerary rationality, service fit, and resource suitability. Based on these indicators, the adjusted itinerary, service content, and resource allocation were scored. The scores of each item were then weighted according to preset weights to obtain the overall satisfaction parameter.

[0139] Step S1645: When the overall satisfaction parameter is greater than the preset satisfaction threshold, it is determined that the optimized preliminary order matching scheme meets the requirements; otherwise, the itinerary arrangement, service content and resource configuration are readjusted until the overall satisfaction parameter is greater than the preset satisfaction threshold.

[0140] The overall satisfaction parameter is compared with the preset satisfaction threshold. If the overall satisfaction parameter is greater than the threshold, it means that the optimized preliminary order matching plan meets the requirements. If it is less than the threshold, the process returns to steps S1641 to S1643 to readjust the itinerary, service content and resource configuration, and conducts another comprehensive evaluation until the overall satisfaction parameter is greater than the preset satisfaction threshold.

[0141] Step S165: Determine the optimized preliminary order matching scheme as the final order matching scheme, format the order matching scheme, and generate order matching scheme display information that meets the user interface display requirements.

[0142] The preliminary order matching plan, optimized through comprehensive evaluation, will be determined as the final order matching plan. This plan will then be formatted, including adjusting the font, size, and layout of the text, adding necessary icons and separators, etc., to meet the display requirements of the travel service platform's user interface, and generating the order matching plan display information.

[0143] Step S166: Send the order matching scheme display information to the user interface of the tourism service platform, and control the user interface to display the order matching scheme display information in a preset display mode.

[0144] The order matching solution information is sent to the user interface of the travel service platform via a data transmission protocol. The user interface presents the order matching solution information to the user in a preset display method, such as pagination or module display, allowing the user to clearly and intuitively view the various contents of the order matching solution.

[0145] For example, the method may also include: Step S210: Collect the user preference feature set, the service provider capability feature set, the demand capability correlation matrix, the order service matching degree parameters, and the corresponding order matching scheme to form a historical order matching data set.

[0146] After each order matching is completed, the user preference feature set, service provider capability feature set, demand capability correlation matrix, order service matching degree parameters, and corresponding order matching scheme generated during the matching process are collected and stored in a designated database, gradually accumulating to form a historical order matching data set.

[0147] Step S211: Perform data cleaning on the historical order matching data set, and extract training samples from the cleaned historical order matching data set. The training samples include sample input data and sample label data. The sample input data is the user preference feature set and the service provider capability feature set. The sample label data is the order service matching degree parameter.

[0148] The historical order matching dataset undergoes data cleaning to remove duplicate data, data with excessive missing values, and outlier data. For example, for user preference feature sets with some missing features, if the missing proportion exceeds a preset threshold, the data is removed; order service matching parameters that significantly deviate from the normal range are identified as outlier data and deleted.

[0149] Training samples are extracted from the cleaned historical order matching data set. Each training sample contains sample input data and sample label data. The sample input data consists of a set of user preference features and a set of service provider capability features. The sample label data is the corresponding order service matching degree parameter.

[0150] Step S310: Construct a neural network model containing an input layer, a hidden layer, and an output layer. The input layer is used to receive the sample input data. The hidden layer includes a feature interaction layer and an attention mechanism layer. The feature interaction layer is used to perform interaction feature extraction processing on the user preference feature set and the service provider capability feature set. The attention mechanism layer is used to perform importance weight allocation processing on the interaction features obtained from the interaction feature extraction processing. The output layer is used to output the order service matching degree prediction parameters.

[0151] A neural network model is constructed, consisting of an input layer, hidden layers, and an output layer. The number of neurons in the input layer matches the feature dimensions of the sample input data, enabling it to receive the sample input data and pass it to the hidden layer.

[0152] The hidden layer comprises a feature interaction layer, an attention mechanism layer, and a fully connected layer. The feature interaction layer uses multiple neurons to extract interactive features from the input set of user preference features and the set of service provider capability features. It generates interactive features by combining and transforming features through the connection weights between neurons.

[0153] The attention mechanism layer contains neurons corresponding to the number of interaction features. Each neuron corresponds to one interaction feature. The importance weight of each interaction feature is determined by calculation, and the interaction features are weighted so that the model pays more attention to the important interaction features.

[0154] The fully connected layer consists of multiple neurons that receive features processed by the attention mechanism layer and perform nonlinear transformations through a nonlinear activation function to generate high-level abstract features.

[0155] The output layer contains a neuron that receives high-level abstract features from the fully connected layer and outputs order service matching degree prediction parameters through a linear activation function.

[0156] Step S311: Input the training samples into the neural network model for model training. By adjusting the parameters of the neural network model, the difference between the order service matching degree prediction parameters and the sample label data meets the preset conditions, and a trained order matching prediction model is obtained.

[0157] Step S3111: Divide the training samples into a training set, a validation set, and a test set. The training set is used for model parameter learning, the validation set is used for model hyperparameter tuning, and the test set is used for model performance evaluation.

[0158] The training samples are divided into a training set, a validation set, and a test set according to a preset ratio, such as 7:2:1. The training set contains most of the training samples and is used by the model to learn the parameter patterns during training. The validation set is used to evaluate the model's performance during training and to help adjust the model's hyperparameters, such as the learning rate and the number of neurons in the hidden layer. The test set is used to finally evaluate the model's generalization ability and performance after the model training is completed.

[0159] Step S3112: Input the sample input data of the training set into the input layer of the neural network model. The feature interaction layer of the hidden layer performs element-wise multiplication and addition operations on the user preference feature set and the service provider capability feature set to generate preliminary interaction features. Then, the attention mechanism layer of the hidden layer calculates the attention weight value of each feature in the preliminary interaction features. The preliminary interaction features are weighted and summed according to the attention weight values ​​to obtain weighted interaction features.

[0160] The sample input data from the training set is fed into the input layer of the neural network model. The input layer then passes the sample input data to the feature interaction layer of the hidden layer. After receiving the user preference feature set and the service provider capability feature set, the feature interaction layer performs element-wise multiplication on the corresponding features in the two feature sets to obtain a set of product features. At the same time, it performs element-wise addition on the corresponding features in the two feature sets to obtain a set of sum features. The product features and the sum features are combined to form the preliminary interaction features.

[0161] The initial interaction features are passed to the attention mechanism layer. The attention mechanism layer determines the attention weight value of each initial interaction feature by calculating its correlation with the overall feature set, and the sum of the attention weight values ​​is 1. Subsequently, based on the attention weight value of each initial interaction feature, the initial interaction features are weighted and summed to obtain weighted interaction features. The weighted interaction features highlight the more important initial interaction feature information.

[0162] Step S3113: Input the weighted interactive features into the fully connected layer of the hidden layer for nonlinear transformation processing to generate high-level abstract features. Input the high-level abstract features into the output layer and output the order service matching degree prediction parameters through the linear activation function of the output layer.

[0163] Weighted interactive features are input into the fully connected layer of the hidden layer. Each neuron in the fully connected layer is connected to all neurons in the previous layer. The weighted interactive features are linearly combined through preset connection weights, and then nonlinear transformation is performed through a nonlinear activation function (such as the ReLU function) to map the low-dimensional features to a high-dimensional space, generating high-level abstract features. High-level abstract features can more profoundly express the complex relationships between features.

[0164] High-level abstract features are input to the output layer, which processes the high-level abstract features through a linear activation function and converts them into order service matching degree prediction parameters. These order service matching degree prediction parameters reflect the model's prediction results on the matching degree between user order demand information and service provision information in the training samples.

[0165] Step S3114: Calculate the loss value between the order service matching degree prediction parameter and the sample label data of the training set. The loss value is calculated using the mean squared error loss function.

[0166] Obtain the sample label data from the training set, i.e., the corresponding order service matching degree parameters. Compare the order service matching degree prediction parameters output by the neural network model with the sample label data, and calculate the loss value between the two using the mean squared error loss function. The mean squared error loss function is calculated as follows: first, calculate the difference between the order service matching degree prediction parameter and the sample label data for each sample; then, square this difference; finally, calculate the average of the squared differences of all samples to obtain the loss value.

[0167] Step S3115: Based on the loss value, the parameters of the neural network model are updated by backpropagation using the gradient descent optimization algorithm, and the weight matrix of the feature interaction layer, the attention weight value of the attention mechanism layer, and the connection weight of the fully connected layer are adjusted.

[0168] Based on the calculated loss value, the gradient descent optimization algorithm is used to adjust the parameters of the neural network model along the direction of decreasing loss. Using backpropagation, the partial derivatives of the loss value with respect to the weight matrix of the feature interaction layer, the attention weights of the attention mechanism layer, and the connection weights of the fully connected layer are calculated; these are the gradients. Based on the direction and magnitude of the gradients, these parameters are updated according to a preset learning rate, gradually reducing the loss value.

[0169] Step S3116: After each preset training round, input the sample input data of the validation set into the currently trained neural network model, calculate the validation order service matching degree prediction parameters, and calculate the validation loss value based on the validation order service matching degree prediction parameters and the sample label data of the validation set.

[0170] After each preset training round, such as every 50 rounds, the sample input data of the validation set is input into the currently trained neural network model. The model processes the data and outputs the validation order service matching degree prediction parameters. The validation order service matching degree prediction parameters are compared with the sample label data of the validation set, and the validation loss value is calculated using the same mean squared error loss function as in step S3114.

[0171] Step S3117: When the verification loss value no longer decreases after a preset number of consecutive steps, stop model training and determine the current neural network model as the trained order matching prediction model.

[0172] Continuously monitor and verify the changes in the loss value. If the loss value stops decreasing after a preset number of times, such as 5 times, it indicates that the model has reached a good training state. Continuing to train may lead to overfitting. At this time, stop the model training and determine the current neural network model as the trained order matching prediction model.

[0173] Step S3118: Use the test set to evaluate the performance of the trained order matching prediction model, calculate the mean absolute error and root mean square error of the model, and determine that the trained order matching prediction model meets the usage requirements when both the mean absolute error and the root mean square error are less than the corresponding preset error threshold.

[0174] Input the sample input data of the test set into the trained order matching prediction model to obtain the test order service matching degree prediction parameters. Calculate the mean absolute error between the test order service matching degree prediction parameters and the sample label data of the test set, which is calculated as the average of the absolute values ​​of the differences between the prediction parameters and the label data for all samples; calculate the root mean square error, which is calculated as the square root of the average of the squares of the differences between the prediction parameters and the label data for all samples.

[0175] The mean absolute error is compared with a preset mean absolute error threshold, and the root mean square error is compared with a preset root mean square error threshold. When both are less than the corresponding preset error threshold, it indicates that the trained order matching prediction model has good prediction performance and generalization ability, and meets the usage requirements.

[0176] Step S312: Deploy the trained order matching prediction model to the order matching system of the tourism service platform to assist in the calculation of order matching degree.

[0177] The trained and usable order matching prediction model is deployed into the order matching system of the tourism service platform. During the order matching process, the user preference feature set and the service provider capability feature set are input into the model. The model outputs the order service matching degree prediction parameters to assist the order matching system in making order matching decisions and improve the efficiency and accuracy of order matching.

[0178] Figure 2 The illustration shows exemplary hardware and software components of an order matching system 100 for a service platform that can implement the ideas of this application, according to some embodiments of this application. For example, a processor 120 can be used in the order matching system 100 of the service platform and to perform the functions described in this application.

[0179] The order matching system 100 of the service platform can be a general-purpose server or a special-purpose server; both can be used to implement the order matching method of the service platform of this application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0180] For example, the order matching system 100 of the service platform may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the order matching system 100 of the service platform may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The order matching system 100 of the service platform also includes an input / output (I / O) interface 150 between the computer and other input / output devices.

[0181] For ease of explanation, only one processor is described in the order matching system 100 of the service platform. However, it should be noted that the order matching system 100 of the service platform in this application may also include multiple processors. Therefore, the steps performed by one processor as described in this application may also be performed jointly or individually by multiple processors. For example, if the processor of the order matching system 100 of the service platform performs steps A and B, it should be understood that steps A and B may also be performed jointly by two different processors or individually by one processor. For example, the first processor performs step A, the second processor performs step B, or the first processor and the second processor jointly perform steps A and B.

[0182] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the order matching method of the service platform described above is implemented.

[0183] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. An order matching method for a service platform, characterized in that, The method includes: The system obtains user order demand information and service provision information from a tourism service platform. The user order demand information includes user itinerary demand information, user service demand information, and user resource demand information. The service provision information includes service provider itinerary provision information, service provider service provision information, and service provider resource provision information. The user order demand information is processed by demand feature extraction to obtain a user preference feature set, which includes trip preference features, service preference features and resource preference features; The service provision information is processed by feature extraction to obtain a service provider capability feature set, which includes trip provision capability features, service provision capability features and resource provision capability features; The user preference feature set and the service provider capability feature set are subjected to feature association processing to generate a demand capability association matrix. Each element in the demand capability association matrix is ​​used to represent the degree of association between a user preference feature in the user preference feature set and a service provider capability feature in the service provider capability feature set. Based on the demand capability association matrix, the matching degree of the user order demand information and the service provision information is calculated to obtain the order service matching degree parameter. The order service matching degree parameter is used to indicate the degree to which the service provision information meets the user order demand information. An order matching scheme is generated based on the order service matching degree parameters, and the order matching scheme is sent to the user interface of the tourism service platform for display. The user order demand information is processed by demand feature extraction to obtain a user preference feature set, which includes trip preference features, service preference features, and resource preference features, including: The user's travel demand information is processed by travel feature parsing to extract travel time features, travel location features, and travel type features from the user's travel demand information. The travel time features, travel location features, and travel type features are then combined to form travel preference features. The user service demand information is processed by service feature parsing to extract service type features, service standard features and service method features from the user service demand information. The service type features, service standard features and service method features are then combined to form service preference features. The user resource demand information is processed by resource feature parsing to extract resource type features, resource quantity features and resource quality features from the user resource demand information, and the resource type features, resource quantity features and resource quality features are combined to form resource preference features; The trip preference features, service preference features, and resource preference features are standardized and integrated to form a set of user preference features that include trip dimensions, service dimensions, and resource dimensions. The step of performing feature extraction processing on the service provision information to obtain a set of service provider capability features includes: The service provider's trip provision information is parsed and processed to extract the trip provision time feature, trip provision location feature, and trip provision type feature from the service provider's trip provision information. The trip provision time feature, the trip provision location feature, and the trip provision type feature are combined to form the trip provision capability feature. The service provider information is parsed and processed to extract service provision type features, service provision standard features, and service provision method features from the service provider information. The service provision type features, service provision standard features, and service provision method features are then combined to form service provision capability features. The resource provision information of the service provider is parsed and processed to extract the resource provision type characteristics, resource provision quantity characteristics, and resource provision quality characteristics from the resource provision information. The resource provision type characteristics, resource provision quantity characteristics, and resource provision quality characteristics are then combined to form resource provision capability characteristics. The trip provision capability characteristics, service provision capability characteristics, and resource provision capability characteristics are standardized and integrated to form a set of service provider capability characteristics that includes trip provision dimension, service provision dimension, and resource provision dimension; The method further includes: Collect the user preference feature set, the service provider capability feature set, the demand capability correlation matrix, the order service matching degree parameters, and the corresponding order matching scheme to form a historical order matching data set; The historical order matching dataset is cleaned, and training samples are extracted from the cleaned historical order matching dataset. The training samples include sample input data and sample label data. The sample input data is the user preference feature set and the service provider capability feature set. The sample label data is the order service matching degree parameter. A neural network model is constructed, comprising an input layer, a hidden layer, and an output layer. The input layer is used to receive the sample input data. The hidden layer includes a feature interaction layer and an attention mechanism layer. The feature interaction layer is used to perform interaction feature extraction processing on the user preference feature set and the service provider capability feature set. The attention mechanism layer is used to perform importance weight allocation processing on the interaction features obtained from the interaction feature extraction processing. The output layer is used to output the order service matching degree prediction parameters. The training samples are input into the neural network model for model training. By adjusting the parameters of the neural network model, the difference between the order service matching prediction parameters and the sample label data meets preset conditions, resulting in a trained order matching prediction model. Specifically, the training samples are divided into a training set, a validation set, and a test set. The training set is used for model parameter learning, the validation set is used for model hyperparameter tuning, and the test set is used for model performance evaluation. The sample input data from the training set is input into the input layer of the neural network model. After passing through the feature interaction layer of the hidden layer, the user preference feature set and the service provider capability feature set are multiplied and added element-wise to generate preliminary interaction features. Then, the attention mechanism layer of the hidden layer calculates the attention weight value of each feature in the preliminary interaction features. The preliminary interaction features are weighted and summed according to the attention weight values ​​to obtain weighted interaction features. The weighted interaction features are input into the fully connected layer of the hidden layer for nonlinear transformation processing to generate high-level abstract features. The high-level abstract features are input into the output layer, and the order matching prediction model is output through the linear activation function of the output layer. The system calculates single-service matching degree prediction parameters; it also calculates the loss value between the order service matching degree prediction parameters and the sample label data of the training set, using the mean squared error loss function; based on the loss value, it uses a gradient descent optimization algorithm to backpropagate and update the parameters of the neural network model, adjusting the weight matrix of the feature interaction layer, the attention weight value of the attention mechanism layer, and the connection weight of the fully connected layer; after each preset training round, it inputs the sample input data of the validation set into the currently trained neural network model, calculates the validation order service matching degree prediction parameters, and calculates the validation loss value based on the validation order service matching degree prediction parameters and the sample label data of the validation set; when the validation loss value no longer decreases for a preset number of consecutive times, it stops model training and determines the current neural network model as the trained order matching prediction model; it uses the test set to evaluate the performance of the trained order matching prediction model, calculates the model's mean absolute error and root mean square error, and determines that the trained order matching prediction model meets the usage requirements when both the mean absolute error and the root mean square error are less than the corresponding preset error thresholds; The trained order matching prediction model is deployed to the order matching system of the tourism service platform. During the order matching process, the user preference feature set and the service provider capability feature set are input into the model, and the order service matching degree prediction parameters are output to assist the order matching system in making order matching decisions.

2. The order matching method of the service platform according to claim 1, characterized in that, The process of parsing and analyzing the user's travel demand information to extract travel time features, travel location features, and travel type features from the user's travel demand information, and combining the travel time features, travel location features, and travel type features to form travel preference features, including: The user's travel request information is processed by text segmentation, which decomposes the user's travel request information into multiple travel text words; The multiple itinerary text words are subjected to part-of-speech tagging to identify words indicating time, words indicating location, and words indicating type; Extract trip time features from words representing time, including trip start time, trip end time, and trip duration; Extract trip location features from words representing locations, including the trip departure point, the points along the trip route, and the trip destination; Extract trip type features from the vocabulary representing types, including trip theme type, trip activity type, and trip distance type; The trip time features, trip location features, and trip type features are combined according to preset feature combination rules to form trip preference features that include time, location, and type dimensions; The combined travel preference features are subjected to feature verification processing to ensure that the travel preference features can accurately reflect the travel preferences in the user's travel demand information.

3. The order matching method for the service platform according to claim 1, characterized in that, The step of performing feature association processing on the user preference feature set and the service provider capability feature set to generate a demand-capability association matrix includes: Determine the number of user preference features included in the user preference feature set and the number of service provider capability features included in the service provider capability feature set; An initial association matrix is ​​constructed based on the number of user preference features and the number of service provider capability features. The number of rows in the initial association matrix is ​​the same as the number of user preference features, and the number of columns in the initial association matrix is ​​the same as the number of service provider capability features. For each user preference feature in the user preference feature set, the correlation value between the user preference feature and each service provider capability feature in the service provider capability feature set is calculated sequentially. The calculated correlation values ​​are filled into the corresponding positions in the initial correlation matrix to generate a demand-capability correlation matrix containing the correlation values ​​between all user preference features and service provider capability features.

4. The order matching method of the service platform according to claim 3, characterized in that, For each user preference feature in the user preference feature set, the correlation value between the user preference feature and each service provider capability feature in the service provider capability feature set is calculated sequentially, including: Obtain the feature attribute descriptions of the user preference features and the feature attribute descriptions of the service provider capability features; By comparing the feature attribute descriptions of the user preference features and the feature attribute descriptions of the service provider capability features, the attribute matching items and attribute difference items between the two are determined; The attribute matching rate is calculated based on the number of attribute matches and the number of attribute differences. The attribute matching rate is the ratio of the number of attribute matches to the sum of the number of attribute matches and the number of attribute differences. Based on the attribute matching rate and combined with the preset attribute weight coefficient, the basic correlation value between the user preference feature and the service provider capability feature is calculated. The basic correlation value is adjusted by weighting the basic correlation value according to the importance coefficients of the user preference features and the service provider capability features to obtain the final correlation value.

5. The order matching method for the service platform according to claim 1, characterized in that, The process of calculating the matching degree between the user order demand information and the service provision information based on the demand capability association matrix to obtain the order service matching degree parameter includes: Obtain the preference weight value of each user preference feature in the user preference feature set and the capability weight value of each service provider capability feature in the service provider capability feature set; Based on the preference weight value and the capability weight value, the correlation degree value in the demand capability correlation matrix is ​​weighted to obtain a weighted demand capability correlation matrix; The weighted demand capability association matrix is ​​summed in the row direction to obtain the total association value corresponding to each user preference feature; The weighted demand-capability correlation matrix is ​​summed along the column direction to obtain the total correlation value corresponding to each service provider's capability feature. The total correlation values ​​corresponding to all user preference features are summed to obtain the total correlation value of user preferences. The total correlation values ​​corresponding to all service provider capability features are summed to obtain the total correlation value of service provider capabilities. Based on the total correlation value of user preferences and the total correlation value of service provider capabilities, an order service matching parameter is calculated. The order service matching parameter is the ratio of the product of the total correlation value of user preferences and the total correlation value of service provider capabilities to the sum of the two.

6. The order matching method of the service platform according to claim 1, characterized in that, The step of generating an order matching scheme based on the order service matching degree parameter and sending the order matching scheme to the user interface of the tourism service platform for display includes: Obtain a preset order matching threshold, and compare the order service matching degree parameter with the order matching threshold; When the order service matching degree parameter is greater than or equal to the order matching threshold, it is determined that the service provision information and the user order demand information are successfully matched, and a preliminary order matching scheme is generated based on the service provision information. When the order service matching degree parameter is less than the order matching threshold, it is determined that the service provision information and the user order demand information fail to match, and the process returns to the step of obtaining service provision information, re-obtaining new service provision information and recalculating the matching degree. The preliminary order matching scheme is optimized by adjusting the itinerary, service content, and resource allocation based on the user preference feature set and the service provider capability feature set. The optimized preliminary order matching scheme is determined as the final order matching scheme. The order matching scheme is then formatted to generate order matching scheme display information that meets the user interface display requirements. The order matching scheme display information is sent to the user interface of the tourism service platform, and the user interface is controlled to display the order matching scheme display information in a preset display mode.

7. The order matching method for the service platform according to claim 6, characterized in that, The optimization process for the preliminary order matching scheme, based on the user preference feature set and the service provider capability feature set, adjusts the itinerary arrangement, service content, and resource allocation in the preliminary order matching scheme, including: Based on the trip preference characteristics and the trip provision capability characteristics, the trip arrangements in the preliminary order matching scheme are adjusted to optimize the trip time sequence and trip location connections; Based on the service preference characteristics and the service provision capability characteristics, the service content in the preliminary order matching scheme is adjusted to increase service items that meet the user's service needs and reduce service items that do not meet the user's service needs. Based on the resource preference characteristics and the resource provision capability characteristics, the resource configuration in the preliminary order matching scheme is adjusted to optimize the resource allocation quantity and resource usage order; A comprehensive evaluation of the adjusted itinerary, service content, and resource allocation was conducted, and the overall satisfaction parameter after the plan was adjusted was calculated. When the overall satisfaction parameter is greater than the preset satisfaction threshold, the optimized preliminary order matching plan is determined to meet the requirements; otherwise, the itinerary arrangement, service content, and resource allocation are readjusted until the overall satisfaction parameter is greater than the preset satisfaction threshold.

8. An order matching system for a service platform, characterized in that, The order matching system of the service platform includes a processor and a memory, the memory and the processor are connected, the memory is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the memory to implement the order matching method of the service platform according to any one of claims 1-7.

Images