A candidate item priority score adjustment method, device and medium for civil aviation

Abstract

The application provides a civil aviation-oriented candidate item priority score adjustment method, device and medium, relating to the technical field of data processing, wherein if the first target user data corresponding to the influence factor meets the matching condition corresponding to the influence factor, the influence factor is taken as an associated factor; the target push score of the candidate item is obtained based on the associated factor to obtain the priority score corresponding to the candidate item; if the associated factor includes all influence factors corresponding to the candidate adjustment rule, the problem text corresponding to the candidate adjustment rule is input into a large language model to obtain a decision result corresponding to the problem text; if the decision result is a first type of result, the priority score corresponding to the candidate item corresponding to the candidate adjustment rule is adjusted according to the score adjustment mode corresponding to the candidate adjustment rule; the adjustment of the priority score through the large language model realizes the automatic adjustment of the priority score of the candidate item, effectively reduces the manual maintenance cost, and improves the data processing efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a candidate item priority score adjustment method for civil aviation, equipment and medium. BACKGROUND

[0002] The existing civil aviation recommendation system usually sorts the recommendation order of candidate item related information according to the priority score of the candidate item, and preferentially displays the candidate item related information corresponding to the candidate item with a higher priority score, so as to improve the user interaction efficiency and the recommendation effect. In the prior art, the method for determining and adjusting the priority score of the candidate item usually depends on the user data or the personalized behavior data of the user collected in a historical time period to obtain the initial priority score of the candidate item, and then the relevant personnel manually adjusts the initial priority score of part of the candidate items according to the actual demand, experience and strategy to obtain the target priority score of the candidate item. However, the process of manually adjusting the initial priority score of part of the candidate items is tedious and time-consuming, and is easily affected by subjective judgment and lacks objective basis. In addition, maintaining a large number of priority adjustment rules also requires a large amount of human cost. SUMMARY

[0003] In view of the above technical problems, the technical scheme adopted by the present application is as follows:

[0004] According to a first aspect of the present application, a candidate item priority score adjustment method for civil aviation is provided, which comprises the following steps:

[0005] S1, for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor corresponding to the target user Y to obtain the associated factor set F corresponding to Y.

[0006] S2, obtaining the target push score matrix D corresponding to Y, the element on the ith row and jth column in D is the target push score of the ith influencing factor to the jth candidate item; 1≤i≤m; 1≤j≤n; m is the number of influencing factors; n is the number of candidate items.

[0007] S3, for each candidate item, the sum of the target push scores of all associated factors in F to the candidate item is taken as the priority score corresponding to the candidate item.

[0008] S4, for each candidate adjustment rule in a plurality of candidate adjustment rules, obtaining the data set corresponding to the candidate adjustment rule, the data set comprising a question text, at least one influencing factor, at least one candidate item and a score adjustment mode.

[0009] S5, for each candidate adjustment rule, if all influencing factors corresponding to the candidate adjustment rule are included in F, input a question text corresponding to the candidate adjustment rule into the large language model to obtain a decision result J corresponding to the question text.

[0010] S6, if J is a first type of result, adjust priority score values corresponding to all candidate items corresponding to the candidate adjustment rule according to a score adjustment mode corresponding to the candidate adjustment rule corresponding to J to obtain a target priority score value corresponding to each candidate item; the first type of result indicates that the priority of the candidate item corresponding to the candidate adjustment rule needs to be adjusted.

[0011] According to a second aspect of the present application, a non-transitory computer readable storage medium is provided, the storage medium storing a computer program, the computer program being loaded and executed by a processor to implement the foregoing method.

[0012] According to a third aspect of the present application, an electronic device is provided, comprising a processor, a memory, and a computer program stored on the memory and executable on the processor, the processor implementing the foregoing method when executing the computer program.

[0013] The present application has at least the following beneficial effects:

[0014] The embodiment of the application provides a candidate item priority score adjustment method for civil aviation, equipment and medium, in the method, for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor corresponding to the target user to obtain the associated factor set corresponding to the target user; a target push score matrix corresponding to the target user is obtained; for each candidate item, the sum of the target push scores of all associated factors in the associated factor set on the candidate item is taken as the priority score corresponding to the candidate item; a data set corresponding to the candidate adjustment rule is obtained, the data set includes a question text, at least one influencing factor, at least one candidate item and a score adjustment mode; for each candidate adjustment rule, if all influencing factors corresponding to the candidate adjustment rule are included in the associated factor set, the question text corresponding to the candidate adjustment rule is input into a large language model to obtain a decision result corresponding to the question text; if the decision result is a first type of result, the priority scores corresponding to all candidate items corresponding to the candidate adjustment rule are adjusted according to the score adjustment mode corresponding to the candidate adjustment rule corresponding to the decision result, to obtain a target priority score corresponding to each candidate item. It can be known that the application only takes the influencing factor as the associated factor when the first target user data meets the matching condition corresponding to the influencing factor, so that the interference of irrelevant factors can be avoided; the priority score corresponding to the candidate item is obtained based on the target push score of the associated factor on the candidate item, the adjustment of the priority score is realized through the large language model, the automatic and intelligent adjustment of the priority score of the candidate item is realized, the accuracy of the target priority score is improved, the artificial maintenance cost is effectively reduced, and the data processing efficiency is significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can be obtained without creative labor.

[0016] Figure 1 A flowchart of a candidate item priority score adjustment method for civil aviation provided by the embodiment of the application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the application will be described clearly and completely in the following with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only some embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0018] It is to be understood that the terms "first", "second", and the like, used in the description and the claims of the application and the above drawings, are used to distinguish between similar objects, and are not necessarily used to describe a particular sequential or chronological order. It is to be understood that the use of data so designated is not to be construed as limiting of the embodiments of the application described herein to only those embodiments exemplified in the figures or described in the specification. Further, the terms "comprises", "comprising", "includes", "including" and the like can be used in the specification to mean one or more of the listed steps or elements included in the process, method, system, product or server, and that is not meant to exclude other steps or elements not specifically listed or inherent to such process, method, system, product or server.

[0019] Embodiments of the application provide a candidate item priority score adjustment method for civil aviation, which comprises the following steps, as shown in the following: Figure 1

[0020] S1, for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor corresponding to the target user Y to obtain the associated factor set F corresponding to Y.

[0021] Specifically, the influencing factor is a factor that affects the push order of related information of the candidate item, for example: hotel preference, car preference, bad weather, browsing preference, ticket preference, evening rush hour, flight itinerary containing car top 10 cities, etc.

[0022] Specifically, the candidate item is an item that can be recommended or displayed to the user, for example: commodity, article, video, service, etc.

[0023] Specifically, the first target user data is obtained based on the data acquisition rule corresponding to the matching condition corresponding to the influencing factor.

[0024] Specifically, the first target user data is obtained based on the data acquisition rule corresponding to the matching condition corresponding to the influencing factor, specifically including:

[0025] For each influencing factor, the first target user data corresponding to the influencing factor is obtained from all data related to Y according to the data acquisition rule corresponding to the matching condition corresponding to the influencing factor.

[0026] ​Specifically, the influencing factor, the matching condition corresponding to the influencing factor, and the data acquisition rule corresponding to the matching condition are stored in a first data configuration table corresponding to the influencing factor. The first data configuration table is a configuration table pre-set by a person skilled in the art according to actual needs, and includes each influencing factor, the matching condition corresponding to each influencing factor, and the data acquisition rule corresponding to each matching condition. For example, the influencing factor is adverse conditions, the matching condition corresponding to the influencing factor is that the current weather type is any one of snow, freezing, low temperature, gale, blowing sand, and heavy rain, and the data acquisition rule corresponding to the matching condition is to acquire the weather type of the arrival place of the current trip at the trip arrival time point.

[0027] In the embodiment, not all influencing factors will affect the priority of the candidate. For example, the influencing factor is adverse weather, but the departure place, arrival place, and passing place of the current trip of the target user do not have adverse weather, so the influencing factor will not affect the priority of the candidate. Through the above steps, only when the first target user data satisfies the matching condition corresponding to the influencing factor, the influencing factor is used as the associated factor, and the interference of irrelevant factors can be avoided.

[0028] S2, acquiring a target push score matrix D corresponding to Y, an element on the ith row and the jth column in D is a target push score of an ith influencing factor to a jth candidate; 1≤i≤m; 1≤j≤n; m is the number of influencing factors; n is the number of candidates.

[0029] S3, for each candidate, adding the target push scores of all associated factors in F to the candidate to obtain a sum as a priority score corresponding to the candidate.

[0030] In a specific embodiment, the greater the priority score of the candidate, the higher the candidate priority corresponding to the candidate.

[0031] Further, the candidate priority is used to determine the push order of the related information of the candidate. The related information of the candidate with a high candidate priority is preferentially pushed to M.

[0032] S4, for each candidate adjustment rule in a plurality of candidate adjustment rules, acquiring a data set corresponding to the candidate adjustment rule, the data set including a question text, at least one influencing factor, at least one candidate, and a score adjustment mode.

[0033] Specifically, the candidate adjustment rule is pre-determined by a person skilled in the art according to actual needs, for example, a ride service enhancement recommendation rule.

[0034] Specifically, the question text includes: a query text; the query text is a text pre-set by a person skilled in the art to clarify the decision-making objective according to the candidate adjustment rule, for example: the candidate adjustment rule is a car service enhancement recommendation rule, and the query text corresponding to the candidate adjustment rule includes the question: whether to adjust the priority score of the candidate options related to car service.

[0035] Specifically, the question text also includes a prompt text, which is used to guide the large language model to obtain information related to the decision-making objective corresponding to the question text. This helps the large language model understand which external data should be collected for auxiliary judgment. For example, based on factors such as the current user's geographical location, local car use policies, and whether there are major events or holidays, please comprehensively judge whether it is necessary to increase the priority score of the candidate options related to car use services.

[0036] Specifically, after receiving the question text, the large language model automatically calls external data sources, such as weather APIs, map APIs, and local databases, to integrate and reason about the information and output the decision result corresponding to the question text.

[0037] Furthermore, the correspondence between the candidate adjustment rules and the influencing factors, as well as the correspondence between the candidate adjustment rules and the candidate options, are predetermined by those skilled in the art based on actual needs. For example, the candidate adjustment rule is a car service enhancement recommendation rule, and the influencing factors corresponding to the candidate adjustment rule are flight itineraries including the top 10 cities for car services and evening peak hours; the candidate options corresponding to the candidate adjustment rule are vehicle reservation services, real-time traffic information push services, and carpooling services.

[0038] Furthermore, the score adjustment method corresponding to the candidate adjustment rule is an adjustment method pre-set by those skilled in the art based on the candidate adjustment rule. For example, if the candidate adjustment rule is a car service enhancement recommendation rule, the score adjustment method corresponding to the candidate adjustment rule is: adjust the priority score to 100; adjust the priority score to 90; and so on.

[0039] S5. For each candidate adjustment rule, if F includes all influencing factors corresponding to the candidate adjustment rule, then input the question text corresponding to the candidate adjustment rule into the large language model to obtain the decision result J corresponding to the question text.

[0040] S6. If J is a first-class result, then the priority scores of all candidates corresponding to the candidate adjustment rule are adjusted according to the score adjustment method corresponding to the candidate adjustment rule corresponding to J, so as to obtain the target priority score corresponding to each candidate; the first-class result indicates that the priority of the candidate corresponding to the candidate adjustment rule needs to be adjusted.

[0041] Specifically, the decision result further includes a second type of result, and the second type of result indicates that the priority of the candidate item corresponding to the candidate adjustment rule does not need to be adjusted.

[0042] Specifically, the greater the target priority score of the candidate item, the higher the priority of the candidate item corresponding to the candidate item.

[0043] Through the above steps, for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor corresponding to the target user to obtain the associated factor set corresponding to the target user; a target push score matrix corresponding to the target user is obtained; for each candidate item, the sum of the target push scores of all associated factors in the associated factor set on the candidate item is taken as the priority score corresponding to the candidate item; a data set corresponding to the candidate adjustment rule is obtained, and the data set includes a question text, at least one influencing factor, at least one candidate item, and a score adjustment mode; for each candidate adjustment rule, if all influencing factors corresponding to the candidate adjustment rule are included in the associated factor set, the question text corresponding to the candidate adjustment rule is input into the large language model to obtain a decision result corresponding to the question text; if the decision result is a first type of result, the priority scores corresponding to all candidate items corresponding to the candidate adjustment rule are adjusted according to the score adjustment mode corresponding to the candidate adjustment rule corresponding to the decision result, to obtain a target priority score corresponding to each candidate item; only when the first target user data meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor, which can avoid the interference of irrelevant factors; the priority score corresponding to the candidate item is obtained based on the target push score of the associated factor on the candidate item, the adjustment of the priority score is realized through the large language model, the automatic and intelligent adjustment of the priority score of the candidate item is realized, the accuracy of the target priority score is improved, the artificial maintenance cost is effectively reduced, and the data processing efficiency is significantly improved.

[0044] Specifically, step S2 includes the following substeps:

[0045] S21, obtaining a basic push score matrix A according to a historical data set L; an element a ij is the basic push score of the i th influencing factor X i on the j th candidate item H j ; L includes all data related to each candidate item collected in a historical time period; for example: clicks, views, collections, purchases, ratings, searches, behavior occurrence times, geographic locations, device types, network environments, ages, genders, occupations, likes, collections, etc.

[0046] Specifically, -10≤a ij ≤10.

[0047] Further, a ij The greater, X i The greater the impact on H j .

[0048] Specifically, the end time point of the historical time period is earlier than the current time point, and the length of the historical time period is a preset first length.

[0049] Specifically, step S21 includes the following steps S211-S212:

[0050] S211, input the historical data set into the gth preset recommendation model to obtain X i The initial push score W j of H g ij , 1≤g≤h, h is the number of preset recommendation models.

[0051] Specifically, -10≤W g ij ≤10.

[0052] Specifically, the preset recommendation model is a recommendation model determined in advance from existing recommendation models, and different preset recommendation models use different recommendation algorithms, for example: a content-based preset recommendation model, a collaborative filtering-based recommendation model, and a rule-based recommendation model.

[0053] S212, based on W g ij Obtain a ij , a ij meets the following conditions:

[0054] a ij =∑ h g=1 (V g ×W g ij ), V g is a preset importance weight corresponding to the gth preset recommendation model.

[0055] Specifically, ∑ h g=1 V g =1.

[0056] Through the above steps, the initial push score is obtained according to the historical data set and the preset recommendation model, the base push score is obtained based on the preset importance weight corresponding to the recommendation model and the initial push score, and the multi-model fusion avoids the deviation of a single model, which is beneficial to improve the accuracy of the obtained base push score.

[0057] In a specific embodiment, the base pushing score of each influencing factor on each candidate can be obtained by other manners, for example, the base pushing score of each influencing factor on each candidate is set by an expert according to experience, system prior knowledge or expert knowledge.

[0058] S22, obtaining a score adjustment weight matrix C corresponding to Y according to the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item and the user label corresponding to Y; an element c ij for a ij corresponding score adjustment weight; the second target user data set includes n second target user data corresponding to the candidate.

[0059] Specifically, the second target user data is obtained based on a data acquisition rule corresponding to the preset behavior item.

[0060] Specifically, 0≤c ij ≤10.

[0061] Specifically, the preset behavior item and the judgment condition corresponding to the preset behavior item are behavior items and judgment conditions pre-set by a person skilled in the art according to actual needs, for example, the preset behavior item is exposure, and the judgment condition corresponding to the preset behavior item is that exposure ten times is not clicked, exposure five times is not clicked; or the preset behavior item is a first exposure click ratio and a second exposure click ratio, the judgment condition corresponding to the first exposure click ratio is that the ratio of the number of exposures to the number of clicks is 0.3-0.5, and the judgment condition corresponding to the second exposure click ratio is that the ratio of the number of exposures to the number of clicks is 0.6-0.8.

[0062] Further, the number of preset behavior items is not less than 1.

[0063] Specifically, the number of judgment conditions corresponding to the preset behavior item is not less than 1.

[0064] Specifically, the number of user labels corresponding to Y is not less than 1.

[0065] Specifically, the second target user data is obtained based on a data acquisition rule corresponding to the preset behavior item, and specifically includes the following steps S01-S02:

[0066] S01, for each preset behavior item, obtaining a data acquisition time period corresponding to the preset behavior item, the end time point of the data acquisition time period is a current time point, and the length of the data acquisition time period is a preset length corresponding to the preset behavior item.

[0067] Preferably, the end time point of the historical time period is earlier than the starting time point of the data acquisition time period corresponding to any one preset behavior item.

[0068] S02、according to the data acquisition rule corresponding to the preset behavior item, collecting the data corresponding to the preset behavior item from the data acquisition time period corresponding to the preset behavior item j All data with an association, obtaining the second target user data corresponding to the preset behavior item in the second target user data set j Corresponding to H.

[0069] Specifically, the data acquisition rule corresponding to the preset behavior item is stored in the second data configuration table corresponding to the preset behavior item. The second data configuration table is a configuration table pre-set by a person skilled in the art according to actual needs, including each preset behavior item and the data acquisition rule corresponding to each preset behavior item.

[0070] Through the above steps, the data acquisition time period corresponding to each preset behavior item is obtained, and the second target user data corresponding to the preset behavior item is obtained based on the data acquisition time period and the data acquisition rule corresponding to the preset behavior item, which can filter out irrelevant or invalid data and avoid processing irrelevant or invalid data, thereby saving resources.

[0071] S23, according to A and C, obtaining D, wherein the element d ij meets the following conditions:

[0072] d ij =a ij ×c ij .

[0073] Specifically, the candidate priority obtained by statistical analysis based on the user behavior data of all users in the historical time period has globality and can reflect the overall popular trend, but ignores the individual differences between users, making it difficult to achieve personalized recommendation. Based on the personalized historical behavior data of the target user, the preference intensity of the target user for each candidate item is predicted through collaborative filtering or deep learning model, which can generate more personalized recommendation results. However, due to the sparsity of the personalized behavior data of the target user, the difference in the priority generated by different users is small, which affects the discrimination and accuracy of the recommendation results.

[0074] By the above steps, first, the basic push score matrix is obtained according to the historical data set; for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor corresponding to the target user to obtain the associated factor set corresponding to the target user; the influence of irrelevant influencing factors is avoided; the score adjustment weight matrix corresponding to the target user is obtained according to the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item, and the user label corresponding to the target user; the target push score matrix corresponding to the target user is obtained according to the basic push score matrix and the score adjustment weight matrix corresponding to the target user, and the target push score matrix includes the target push score of each influencing factor to each candidate item; the basic push score matrix is modified in combination with the score adjustment weight matrix corresponding to the target user, and a dynamic weight matrix and a multi-source data fusion mechanism are used to obtain the target push score matrix corresponding to the target user; it is beneficial to improve the discrimination and accuracy of the target push score matrix; further, the sum of the target push scores of all associated factors in the associated factor set to the candidate item is taken as the priority score corresponding to the candidate item; while retaining the global statistical advantage, the personalization degree and the discrimination of the push are significantly improved, the discrimination and accuracy of the priority of the candidate item corresponding to different users are enhanced, and further, it is beneficial to improve the discrimination and accuracy of the recommendation result.

[0075] Specifically, step S22 includes the following steps:

[0076] S221, for each preset behavior item, if the second target user data corresponding to the preset behavior item meets the judgment condition corresponding to the preset behavior item, the basic adjustment weight corresponding to each influencing factor of H j to the basic push score is updated according to the weight update rule corresponding to the judgment condition, so that E is updated; E is a basic adjustment matrix, the element on the i-th row and the j-th column in E is a j corresponding basic adjustment weight, and the initial value of the basic adjustment weight is 1. ij S222, for each preset behavior item, if the second target user data corresponding to the preset behavior item does not meet the judgment condition corresponding to the preset behavior item, the basic adjustment weight corresponding to each influencing factor of H j to the basic push score is updated according to the weight update rule corresponding to the judgment condition, so that E is updated; E is a basic adjustment matrix, the element on the i-th row and the j-th column in E is a j corresponding basic adjustment weight, and the initial value of the basic adjustment weight is 1.

[0077] Specifically, the weight update rule corresponding to the judgment condition is an update rule pre-set by a person skilled in the art according to actual needs, for example: multiply 0.2; multiply 6; multiply 3; multiply 0.5; add 2; add 5, subtract 2; and details are not described here.

[0078] Specifically, each update of the basic adjustment weight is based on the latest basic adjustment weight obtained from the previous update. According to the priority or execution order of the preset behavior items, the basic adjustment weight corresponding to the basic push score of the candidate item is iteratively adjusted for each influencing factor corresponding to the preset behavior item in turn, so as to ensure that the update process has time sequence and continuity. The priority or execution order of the preset behavior items is determined in advance by those skilled in the art according to actual needs, and will not be elaborated here.

[0079] Specifically, the correspondence between preset behavioral items and influencing factors is determined in advance by those skilled in the art based on actual needs, and will not be elaborated here.

[0080] S222. Use the updated E as the first adjustment matrix G corresponding to M; the element in the i-th row and j-th column of G is a. ij The corresponding first adjustment weight.

[0081] S223. Based on the data validity judgment conditions corresponding to the preset behavior item, determine whether each second target user data corresponding to the preset behavior item is valid data.

[0082] Specifically, the data validity judgment conditions corresponding to the preset behavior items are conditions pre-set by those skilled in the art based on actual needs to determine whether the data is valid. For example, if the preset behavior item is exposure, the data validity judgment corresponding to the preset behavior item is that the exposure duration is greater than the preset minimum exposure duration. If the data in the second target user data corresponding to the preset behavior item indicates that the exposure duration is not greater than the preset minimum exposure duration, then the second target user data is determined to be invalid data. If the data in the second target user data corresponding to the preset behavior item indicates that the exposure duration is greater than the preset minimum exposure duration, then the second target user data is determined to be valid data.

[0083] S224, Regarding H j If in the second target user dataset corresponding to all preset behavior items, there is a match with H j If all the corresponding second target user data are invalid, then H is determined. j Data label H_G in G j Invalid; otherwise, determine H_G j It is valid.

[0084] S225, According to G, H_G j Get C from the user tag corresponding to Y.

[0085] In the embodiment, when the target user is a new user or the target user unintentionally clicks or swipes some content, partial invalid data can be generated; through the above steps, whether the second target user data corresponding to each preset behavior item is valid data is determined according to the data validity judgment condition corresponding to the preset behavior item, if all the second target user data corresponding to the preset behavior items are invalid data, then the first adjustment weight corresponding to each influence factor on the basis of the push score of the candidate item is also invalid, therefore, it is determined that the data label of the candidate item in the first adjustment weight matrix is invalid, otherwise, it is determined that the data label of the candidate item in the first adjustment weight matrix is valid, the score adjustment weight matrix corresponding to the target user is obtained according to the data label of the candidate item in the first adjustment weight matrix, which is beneficial to improve the accuracy of the score adjustment weight matrix.

[0086] Specifically, the step S225 comprises the following steps:

[0087] S100, when at least α candidate items in G have valid data labels and not all candidate items in G have valid data labels, if H_G j is invalid, it is determined that H j is the first candidate item; if H_G j is valid, it is determined that H j is the second candidate item; to obtain the first candidate item list L=(L1, L2, …, L k , …, L t ) and the second candidate item list P=(P1, P2, …, P y , …, P q ); α is a preset number of valid items; L k is the kth first candidate item, 1≤k≤t, t is the number of first candidate items, P y is the yth second candidate item, 1≤y≤q, q is the number of second candidate items.

[0088] Specifically, 1<α<n.

[0089] Specifically, t+q=n.

[0090] Specifically, in the step S100, if each candidate item in G has a valid data label or an invalid data label, G is taken as R.

[0091] S200, from all the first adjustment matrices of all other users obtained in a preset time period, L k corresponding first intermediate matrix list Q k =(Q k1 , Q k2 , …, Qkr , …, Q ks(k) ), Q kr is L k the corresponding rth first intermediate matrix, 1≤r≤s(k), s(k) is L k the number of corresponding first intermediate matrices; P1, P2, …, P y , …, P q and L k all data tags in Q kr are valid; it can be understood that Q kr belongs to a set constructed by all other users corresponding first adjustment matrices obtained within a preset time period.

[0092] Specifically, the other user is any one user except the target user.

[0093] Specifically, the end time point of the preset time period is the current time point, and the duration of the preset time period is a preset second duration, which is less than the preset first duration; preferably, the end time point of the preset time period is the current time point.

[0094] S300, if the similarity between G 0 and Q kr corresponding third intermediate matrix Q 0 kr is greater than a preset similarity threshold S 0 , then Q kr is taken as the first key matrix corresponding to L k to obtain the first key matrix list V k corresponding to L k = (V k1 , V k2 , …, V kz , …, V ku(k) ), V kz is the zth first key matrix corresponding to L k , 1≤z≤u(k), u(k) is L k the number of corresponding first key matrices; G 0 and Q 0 kr are m×q matrices and the element in the ith row and yth column in G 0 is X i in G y the first adjustment weight corresponding to the basis push score of P 0 kr is the element in the ith row and yth column in Q kr is X i in Q yThe first adjustment weight corresponding to the basic push score; those skilled in the art know that any method for obtaining the similarity between two matrices in the prior art is within the protection scope of this invention, and will not be described in detail here.

[0095] Specifically, G 0 With Q 0 kr The greater the similarity between them, the better G 0 With Q 0 kr The more similar they are.

[0096] Specifically, 0.8≤S 0 ≤1.

[0097] S400, V k1 V k2 , ..., V kz , ..., V ku(k) ZhongX i For L k The weighted average of the first adjusted weights corresponding to the basic push score is used as X in G. i For L k The first adjustment weight corresponding to the basic push score is used to update G and then use the updated G as the second adjustment matrix R corresponding to M.

[0098] S500, Update R according to the user tag corresponding to Y to obtain C.

[0099] By the above steps, when there are at least a candidate whose data label in the first adjustment weight matrix corresponding to the target user is valid and not all candidates whose data labels in the first adjustment weight matrix are valid, the first adjustment weight of the target user can be updated according to the first adjustment weight corresponding to the other users, the candidate whose data label in the first adjustment weight matrix corresponding to the target user is invalid is taken as the first candidate, the candidate whose data label in the second adjustment weight matrix corresponding to the target user is valid is taken as the second candidate, for each first candidate, a first intermediate matrix in which the first candidate and all second candidates have valid data labels is screened out from the first adjustment matrix corresponding to the other users, a first key matrix in which the data label of the first candidate is valid and similar to the first adjustment matrix corresponding to the target user is screened out according to the similarity between the second intermediate matrix and the third intermediate matrix, and the first adjustment weight corresponding to the basic push score of the candidate of each influencing factor in the first key matrix is used to iteratively update the first adjustment weight corresponding to the basic push score of the candidate of each influencing factor in the first adjustment weight matrix corresponding to the target user; and the first adjustment weight matrix after all updates is taken as the second adjustment weight matrix corresponding to the target user. The compensation mechanism is used to improve the accuracy of the second adjustment weight matrix, and further, the accuracy of the score adjustment weight matrix is improved.

[0100] Specifically, in step S500, the following steps are further included:

[0101] S501, for each user label corresponding to Y, if the user label is the same as any one of the preset labels, the second adjustment weight corresponding to the basic push score of each candidate of each influencing factor corresponding to the user label is updated according to the weight update rule corresponding to the preset label which is the same as the user label, so as to update R.

[0102] Specifically, the weight update rule corresponding to the preset label is an update rule pre-set by a person skilled in the art according to actual needs, for example: multiplied by 0.2; multiplied by 6; multiplied by 3; multiplied by 0.5; added by 2; added by 5, subtracted by 3; set to 4; set to 0; and details are not described herein.

[0103] Specifically, each update of the second adjustment weight is based on the latest second adjustment weight obtained by the previous update, and the second adjustment weight corresponding to the basic push score of the candidate of each influencing factor corresponding to the user label is iteratively adjusted according to the priority or execution order of the user label, so as to ensure that the update process has time sequence and continuity.

[0104] Specifically, the influencing factors corresponding to the user label are consistent with the influencing factors corresponding to the preset label which is the same as the user label.

[0105] Specifically, the correspondence between the preset label and the influencing factor is preset by a person skilled in the art according to actual needs, which is not described herein.

[0106] S502, taking R after all updates as C.

[0107] Through the above steps, the second adjustment weight matrix is updated multiple times according to the user label of the target user to obtain the score adjustment weight matrix corresponding to the target user, which is beneficial to improve the personalization degree and discrimination degree of the obtained score adjustment weight matrix.

[0108] In one specific embodiment, after step S200 and before step S500, the following step of obtaining R is further included:

[0109] S210, if the similarity between G corresponding to the first specified matrix G 1 and Q kr corresponding to the second specified matrix Q 1 kr is greater than S 0 , then Q kr is taken as L k corresponding to the second key matrix to obtain L k corresponding to the second key matrix list W k =(W k1 , W k2 , …, W kβ , …, W kθ(k) ), W kβ is the βth second key matrix corresponding to L k , 1≤β≤θ(k), θ(k) is the number of second key matrices corresponding to L k , G 1 and Q 1 kr are matrices of f×q and G 1 , the element on the bth row and yth column in G b corresponds to the first adjustment weight of the basic push score of P y , Q 1 kr , the element on the bth row and yth column in Q kr corresponds to the first adjustment weight of the basic push score of P b , f is the number of associated factors in F y , F b is the bth associated factor in F, 1≤b≤f.

[0110] Specifically, the greater the similarity between G 1 and Q 1 kr , the greater the similarity between G1 with Q 1 kr the more similar.

[0111] S220, updating G according to the weighted average value of the first adjustment weight corresponding to the basic push score of the candidate item in the second key matrix and the first adjustment weight corresponding to the basic push score of the candidate item in the first adjustment weight matrix corresponding to the target user, to obtain R. k1 , W k2 , …, W kβ , …, W kθ(k) F in G b the basic push score of L k the first adjustment weight corresponding to the basic push score of L b the basic push score of L k the first adjustment weight corresponding to the basic push score of L

[0112] Through the above steps, since not all influencing factors will affect the priority of the candidate item, according to the similarity between the first specified matrix and the second specified matrix, the second key matrix whose data label of the first candidate item is effective and similar to the first adjustment matrix corresponding to the target user is screened out, and based on the first adjustment weight corresponding to the basic push score of the candidate item in each associated factor in the second key matrix, the first adjustment weight corresponding to the basic push score of the candidate item in each associated factor in the first adjustment weight matrix corresponding to the target user is iteratively updated; and the first adjustment weight matrix after all updates is taken as the second adjustment weight matrix corresponding to the target user; compared with the above embodiment, the amount of data to be processed is reduced, which is conducive to improving the calculation efficiency, and at the same time, the influence of irrelevant influencing factors is reduced, which is conducive to improving the accuracy of the second adjustment weight matrix, and further, it is conducive to improving the accuracy of the score adjustment weight matrix.

[0113] In a specific embodiment, after step S222, the following sub-step S001 is included:

[0114] S001, updating G according to the user label corresponding to Y to obtain C.

[0115] Through the above steps, the first adjustment weight matrix is obtained by updating and iterating the basic adjustment matrix multiple times according to the preset behavior item and the second target user data corresponding to the preset behavior item, and the score adjustment weight matrix corresponding to the target user is obtained by updating the first adjustment weight matrix according to the user label corresponding to the target user; the score adjustment weight matrix is determined according to the characteristics of the target user, which is conducive to improving the individualization degree and discrimination degree of the obtained score adjustment weight matrix.

[0116] Specifically, in step S001, the following sub-steps S0011-S0012 are included:

[0117] S0011、for each user label corresponding to Y, if the user label is the same as any one of the preset labels, then the first adjustment weight corresponding to each influence factor of the user label for each candidate item is updated according to the weight update rule corresponding to the preset label same as the user label, so as to update G.

[0118] Specifically, each update of the first adjustment weight is based on the latest first adjustment weight obtained by the previous update, and the first adjustment weight corresponding to each influence factor of the user label for the basic push score of the candidate item is iteratively adjusted according to the priority or execution order of the user label, so as to ensure that the update process has time sequence and continuity; the priority or execution order of the user label is determined by a person skilled in the art according to actual needs, and will not be described here.

[0119] S0012、the G after all updates is taken as C.

[0120] Through the above steps, the first adjustment weight matrix is updated and iterated according to the user label of the target user to obtain the score adjustment weight matrix corresponding to the target user, which is beneficial to improve the individualization degree and discrimination degree of the obtained score adjustment weight matrix.

[0121] Embodiments of the present application also provide a non-transitory computer readable storage medium, which can be arranged in an electronic device to save a computer program related to a method in the method embodiments, and the computer program is loaded and executed by the processor to realize the method provided by the above embodiments.

[0122] Embodiments of the present application also provide an electronic device, comprising a processor, a memory and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to realize the method provided by the above embodiments.

[0123] Embodiments of the present application also provide a computer program product, which includes program code, and when the program product is executed on an electronic device, the program code is used to make the electronic device execute the steps in the method according to various exemplary embodiments of the present application described in the specification.

[0124] Embodiments of the present application provide a candidate item priority score adjustment method for civil aviation, equipment and medium, in the method, for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor corresponding to the target user to obtain the associated factor set corresponding to the target user; obtain the target push score matrix corresponding to the target user; for each candidate item, the sum of the target push scores of all associated factors in the associated factor set to the candidate item is taken as the priority score corresponding to the candidate item; obtain the data set corresponding to the candidate adjustment rule, the data set includes question text, at least one influencing factor, at least one candidate item and score adjustment mode; for each candidate adjustment rule, if all influencing factors corresponding to the candidate adjustment rule are included in the associated factor set, input the question text corresponding to the candidate adjustment rule into the large language model to obtain the decision result corresponding to the question text; if the decision result is the first type of result, adjust the priority scores corresponding to all candidate items corresponding to the candidate adjustment rule according to the score adjustment mode corresponding to the candidate adjustment rule corresponding to the decision result, to obtain the target priority score corresponding to each candidate item. It can be seen that the present application only takes the influencing factor as the associated factor when the first target user data meets the matching condition corresponding to the influencing factor, which can avoid the interference of irrelevant factors; the priority score corresponding to the candidate item is obtained based on the target push score of the associated factor to the candidate item, the adjustment of the priority score is realized through the large language model, which realizes the automatic and intelligent adjustment of the priority score of the candidate item, not only improves the accuracy of the target priority score, but also effectively reduces the artificial maintenance cost, and significantly improves the data processing efficiency.

[0125] Although some specific embodiments of the present application have been described in detail through examples, those skilled in the art should understand that the above examples are only for illustration, not for limiting the scope of the present application. Those skilled in the art should also understand that various modifications can be made to the embodiments without departing from the scope and spirit of the present application.

Claims

1. A candidate item priority score adjustment method for civil aviation, characterized in that, The method comprises the following steps: S1, for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is taken as the associated factor corresponding to the target user Y to obtain the associated factor set F corresponding to Y; the influencing factor, the matching condition corresponding to the influencing factor and the data acquisition rule corresponding to the matching condition are stored in the first data configuration table corresponding to the influencing factor, and the first data configuration table is a pre-set configuration table; S2, obtaining the target push score matrix D corresponding to Y, wherein the element on the i-th row and the j-th column in D is the target push score of the i-th influencing factor to the j-th candidate; 1≤i≤m; 1≤j≤n; m is the number of influencing factors; n is the number of candidate items; in step S2, the following substeps are included: S21, obtaining a basic push score matrix A according to a historical data set L; an element a ij is a basic push score of the i th influencing factor X i on the j th candidate H j ; L includes all data related to each candidate collected in a historical time period; S22、According to the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item and the user tag corresponding to Y, obtain the score adjustment weight matrix C corresponding to Y; The element c on the i-th row and the j-th column in C ij For a ij The corresponding score adjustment weight; The second target user data set includes n second target user data corresponding to the candidate items one by one; 0 ij ; S23、According to A and C, get D, the element d on the i-th row and j-th column in D ij Fits: d ij = a ij × c ij ; S3, for each candidate item, the sum of the target push scores of all associated factors in F to the candidate item is taken as the priority score corresponding to the candidate item; S4, for each candidate adjustment rule in a plurality of candidate adjustment rules, obtaining a data set corresponding to the candidate adjustment rule, wherein the data set comprises a question text, at least one influencing factor, at least one candidate item and a score adjustment mode; S5, for each candidate adjustment rule, if all influencing factors corresponding to the candidate adjustment rule are included in F, the question text corresponding to the candidate adjustment rule is input into the large language model to obtain a decision result J corresponding to the question text; After receiving the question text, the large language model automatically calls an external data source according to the question text, integrates and reasons information, and outputs a decision result corresponding to the question text; S6, if J is a first type of result, adjusting the priority scores corresponding to all candidate items corresponding to the candidate adjustment rule according to the score adjustment mode corresponding to the candidate adjustment rule to obtain a target priority score corresponding to each candidate item; The first type of result indicates that the priority of the candidate item corresponding to the candidate adjustment rule needs to be adjusted.

2. The method of claim 1, wherein the candidate item priority score adjustment method is for civil aviation. In step S22, the following steps are included: S221、for each preset behavior item, if the second target user data corresponding to the preset behavior item meets the judgment condition corresponding to the preset behavior item, updating the basic adjustment weight corresponding to each influence factor of the preset behavior item according to the weight updating rule corresponding to the judgment condition, so as to update E; E is a basic adjustment matrix, and an element on the i th row and the j th column in E is a j ij j ; the initial value of the basic adjustment weight is 1. ij ​ S222. Use the updated E as the first adjustment matrix G corresponding to M; the element in the i-th row and j-th column of G is a. ij The corresponding first adjustment weight; S223, determining whether each second target user data corresponding to the preset behavior item is valid data according to the data validity judgment condition corresponding to the preset behavior item; S224, Regarding H j If in the second target user dataset corresponding to all preset behavior items, there is a match with H j If all the corresponding second target user data are invalid, then H is determined. j Data label H_G in G j Invalid; otherwise, determine H_G j It is valid; S225、According to G, H_G j and Y corresponds to the user label acquisition C.

3. The method of claim 2, wherein the candidate item priority score adjustment is performed for each of the plurality of candidate items. In step S225, the following steps are included: S100、when there are at least α valid data tags in G and not all data tags in G are valid, if H_G j is invalid, determining H j as the first candidate; if H_G j is valid, determining H j as the second candidate; to obtain a first candidate list L=(L1, L2, …, L k , …, L t ) and a second candidate list P=(P1, P2, …, P y , …, P q ); α is a preset valid item number; L k is the kth first candidate, 1≤k≤t, t is the number of first candidates, P y is the yth second candidate, 1≤y≤q, q is the number of second candidates; S200、From all the first adjustment matrices corresponding to all other users obtained within the preset time period, determine L k corresponding first intermediate matrix list Q k =(Q k1 , Q k2 , …, Q kr , … Q ks(k) ), Q kr is the L k corresponding rth first intermediate matrix, 1≤r≤s(k), s(k) is the L k corresponding first intermediate matrix; P1, P2, …, P y , …, P q and L k The data tags in Q kr are all valid; S300, if the similarity between G corresponding to the second intermediate matrix G 0 and Q kr corresponding to the third intermediate matrix Q 0 is greater than a preset similarity threshold S kr , then Q 0 is taken as L kr corresponding to the first key matrix to obtain L k corresponding to the first key matrix list V k = (V k , V k1 , …, V k2 , …, V kz , …, V ku(k) ), V kz is the zth first key matrix corresponding to L k , 1≤z≤u(k), u(k) is the number of first key matrices corresponding to L k ; G 0 and Q 0 kr are matrices of m×q and G 0 is the element in the ith row and yth column of G i , X y is the first adjustment weight corresponding to the basic push score of P 0 ; Q kr is the element in the ith row and yth column of Q kr , X i is the first adjustment weight corresponding to the basic push score of P y ; S400, V k1 V k2 , ..., V kz , ..., V ku(k) ZhongX i For L k The weighted average of the first adjusted weights corresponding to the basic push score is used as X in G. i For L k The first adjustment weight corresponding to the basic push score is used to update G and use the updated G as the second adjustment matrix R corresponding to M. S500, updating R according to the user label corresponding to Y to obtain C.

4. The method of claim 3, wherein the candidate item priority score adjustment method is for civil aviation. After step S200 and before step S500, the following step of obtaining R is further included: S210. If the first specified matrix G corresponds to G... 1 With Q kr The corresponding second specified matrix Q 1 kr The similarity between them is greater than S 0 Then Q kr As L k The corresponding second key matrix is ​​used to obtain L k The corresponding second key matrix list W k =(W k1 W k2 ,…,W kβ ,…,W kθ(k) ), W kβ For L k The corresponding β-th second key matrix, 1≤β≤θ(k), where θ(k) is L k The number of corresponding second key matrices; G 1 and Q 1 kr Both are f×q matrices and G 1 The element in row b and column y of G is F. b For P y The first adjustment weight corresponding to the basic push score, Q 1 kr The element in the b-th row and y-th column is Q. kr China F b For P y The first adjustment weight corresponding to the basic push score, where f is the number of related factors in F, F b Let f be the b-th related factor in F, where 1 ≤ b ≤ f; S220, W k1 , W k2 , …, W kβ , …, W kθ(k) F b the basis push score of L k corresponding to the first adjustment weight of the weighted average value as F b the basis push score of L k corresponding to the first adjustment weight, so that G is updated and the G after all updates are completed as R.

5. The method of claim 3, wherein the candidate item priority score adjustment method is oriented toward civil aviation, and In step S100, if the data label of each candidate item in G is valid, G is taken as R.

6. The method of claim 1, wherein the candidate item priority score adjustment is for civil aviation. -10≤a ij ≤10。 7. The method of claim 1, wherein the candidate item priority score adjustment is for civil aviation. c ij ≤10。 8. A non-transitory computer-readable storage medium, comprising: The storage medium stores a computer program, the computer program is loaded and executed by the processor to realize the candidate item priority score adjustment method for civil aviation in any one of claims 1-7.

9. An electronic device comprising: The processor, the memory and the computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to realize the candidate item priority score adjustment method for civil aviation in any one of claims 1-7. The processor, the memory and the computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to realize the candidate item priority score adjustment method for civil aviation in any one of claims 1-7.

Images