CN115809373B

CN115809373B - Intelligent recommendation method, system and storage medium

Info

Publication number: CN115809373B
Application number: CN202310064012.XA
Authority: CN
Inventors: 赵燚; 杨宇; 田雪彤
Original assignee: Yizhi Technology Co ltd
Current assignee: Yizhi Technology Co ltd
Priority date: 2023-02-06
Filing date: 2023-02-06
Publication date: 2023-06-09
Anticipated expiration: 2043-02-06
Also published as: CN115809373A

Abstract

The embodiment of the specification provides an intelligent recommendation method, an intelligent recommendation system and a storage medium, wherein the method comprises the following steps: the first processing device obtaining a first matrix comprising at least one evaluation index, the at least one evaluation index being determined based on at least one evaluation floating point value, the at least one evaluation floating point value being determined based on at least one recommended object and at least one quantized tag type; the second processing device establishing a second matrix comprising at least one demand indicator, the at least one demand indicator determined based on at least one demand floating point value, the at least one demand floating point value determined based on the at least one recommended subject and the at least one quantized tag type; the third processing equipment acquires a third matrix according to the first matrix and the second matrix; the third processing device determines at least one corresponding target recommended object according to the target recommended subject based on the third matrix.

Description

Intelligent recommendation method, system and storage medium

Technical Field

The present disclosure relates to the field of computer information technologies, and in particular, to an intelligent recommendation method, system, and storage medium.

Background

Recommendation systems are commonly used to match recommended subjects and recommended objects, for example, clients and construction units with construction requirements. However, due to the redundancy of industry information and numerous influencing factors, the requirements of recommended subjects and the business capability of recommended objects are difficult to accurately measure, so that the matching precision is not high.

Therefore, it is necessary to provide an intelligent recommendation scheme, which can improve the matching accuracy of recommendation.

Disclosure of Invention

One or more embodiments of the present disclosure solve the technical problem that the requirements of the recommended subject and the business capabilities of the recommended object are difficult to measure accurately, and improve the accuracy of recommendation matching.

One or more embodiments of the present specification provide an intelligent recommendation method, the method including: the first processing device obtaining a first matrix comprising at least one evaluation index, the at least one evaluation index being determined based on at least one evaluation floating point value, the at least one evaluation floating point value being determined based on at least one recommended object and at least one quantized tag type; the second processing device establishing a second matrix comprising at least one demand indicator, the at least one demand indicator determined based on at least one demand floating point value, the at least one demand floating point value determined based on the at least one recommended subject and the at least one quantized tag type; the third processing equipment acquires a third matrix according to the first matrix and the second matrix; the third processing device determines at least one corresponding target recommended object according to the target recommended subject based on the third matrix.

One or more embodiments of the present specification provide an intelligent recommendation system, the system comprising: a first matrix acquisition module for the first processing device to acquire a first matrix comprising at least one evaluation index, the at least one evaluation index being determined based on at least one evaluation floating point value, the at least one evaluation floating point value being determined based on the at least one recommended object and the at least one quantized tag type; a second matrix building module for the second processing device to build a second matrix comprising at least one demand indicator, the at least one demand indicator determined based on at least one demand floating point value, the at least one demand floating point value determined based on the at least one recommended subject and the at least one quantized tag type; the third matrix acquisition module is used for acquiring a third matrix according to the first matrix and the second matrix by third processing equipment; and the recommendation module is used for determining at least one corresponding target recommended object according to the target recommended subject based on the third matrix by the third processing equipment.

One or more embodiments of the present specification provide a computer-readable storage medium storing computer instructions that, when read by a computer in the storage medium, perform an intelligent recommendation method.

One or more embodiments of the present specification may linearize the evaluation index and the demand index of the recommended object and the recommended subject, respectively, based on the quantization tag type, so that the matching accuracy of the target recommended subject and the target recommended object, which are matched based on the evaluation index and the demand index, is higher.

One or more embodiments of the present disclosure may convert the capability of a recommended object into a linear quantized evaluation floating point value from multiple dimensions through different quantization tags based on experience of the recommended object executing a service, convert a complex recommended object evaluation system into structured data of interest of the service, and improve the accuracy of subsequent evaluation of the recommended object.

According to one or more embodiments of the specification, the range of the evaluation floating point value to be adjusted can be expanded based on actual service information, the problem that the actual situation when a recommended object executes a service and the evaluation floating point value cannot be completely matched is solved, errors caused by the relation between a recommended subject and the recommended object, different performance of the recommended object service, change of industry environment and the like are balanced, meanwhile, the recommended object can be stimulated through a variable expansion range to improve the service capacity corresponding to a quantized tag and the aggressiveness of the executed service, in addition, the evaluation floating point value is corrected through a triggering event, and the evaluation floating point value can be regulated by combining with the evaluation of a third party of the industry, so that the authenticity and the practicability of the evaluation index of the recommended object are improved.

One or more embodiments of the present disclosure determine an evaluation floating point value based on the evaluation floating point value and the first weight, and evaluate the capability of the recommended object from objective experience and subjective willingness of the recommended object to perform a service, so that on one hand, the recommended object having the corresponding capability can be recommended to the recommended subject, on the other hand, the recommended subject can be matched based on the willingness of the recommended object, and meanwhile, the exposure possibility of the rare process of the masses is improved.

Drawings

The present specification will be further elucidated by way of example embodiments, which will be described in detail by means of the accompanying drawings. The embodiments are not limiting, in which like numerals represent like structures, wherein:

FIG. 1 is a schematic illustration of an exemplary application scenario of an intelligent recommendation system, according to some embodiments of the present description;

FIG. 2 is an exemplary block diagram of an intelligent recommendation system, shown in accordance with some embodiments of the present description;

FIG. 3 is an exemplary flow chart of an intelligent recommendation method according to some embodiments of the present description;

FIG. 4 is an exemplary flow chart of a method of determining an evaluation floating point value according to some embodiments of the present disclosure;

FIG. 5 is an exemplary diagram illustrating a method of determining an evaluation floating point value according to some embodiments of the present disclosure.

Detailed Description

In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings that are required to be used in the description of the embodiments will be briefly described below. It is apparent that the drawings in the following description are only some examples or embodiments of the present specification, and it is possible for those of ordinary skill in the art to apply the present specification to other similar situations according to the drawings without inventive effort. Unless otherwise apparent from the context of the language or otherwise specified, like reference numerals in the figures refer to like structures or operations.

It will be appreciated that "system," "apparatus," "unit" and/or "module" as used herein is one method for distinguishing between different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

As used in this specification and the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.

A flowchart is used in this specification to describe the operations performed by the system according to embodiments of the present specification. It should be appreciated that the preceding or following operations are not necessarily performed in order precisely. Rather, the steps may be processed in reverse order or simultaneously. Also, other operations may be added to or removed from these processes.

FIG. 1 is a schematic illustration of an exemplary application scenario of an intelligent recommendation system, according to some embodiments of the present description. In some embodiments, the intelligent recommendation system may be applied to a variety of recommendation scenarios. For example, a commodity of possible interest is recommended to the user. For another example, a construction unit is recommended to a customer having a construction requirement. The intelligent recommendation system 100 may determine the target recommended objects corresponding to the target recommended subjects by implementing the methods and/or processes disclosed herein. For the purpose of illustration, the specification describes a procedure for determining a target recommended object by taking a construction application scenario to recommend a construction unit to a customer. It should be understood that this does not limit the scope of the present application, and the model training method disclosed in this specification can be used to generate target models in any field.

In some embodiments, as shown in fig. 1, the intelligent recommendation system 100 may include a first processing device 110, a second processing device 120, a third processing device 130, a user terminal 140, and a network 150.

The first processing device 110 may be configured to process data and/or information of the recommended objects. For example, the first processing device 110 may determine the corresponding evaluation floating point value based on the number of execution traffic corresponding to the at least one quantized tag type of the recommended object. For another example, the first processing device 110 may determine the corresponding first weight based on an importance score corresponding to at least one quantized tag type of the recommended object. For another example, the first processing device 110 may obtain a first matrix including at least one evaluation index based on the evaluation floating point value and the first weight corresponding to the at least one quantized tag type of the at least one recommended object.

The second processing device 120 may be used to process data and/or information of the recommended subject. For example, the second processing device 120 may determine the corresponding demand floating point value based on the importance scores corresponding to the at least one quantized tag type of the recommended subject. For another example, the second processing device 120 may determine the corresponding second weight based on a tagging behavior of the recommendation body to the at least one quantized tag type corresponding content. For another example, the second processing device 120 may establish a second matrix including at least one demand indicator based on the demand floating point value and the second weight corresponding to the at least one quantized tag type of the at least one recommended subject.

The third processing device 130 may be used to match a recommended subject and a recommended object. For example, the third processing device 130 may determine the target recommendation subject based on the request behavior of the recommendation subject. For another example, the third processing device 130 may acquire a third matrix based on the first matrix and the second matrix, and determine a target recommended object corresponding to the target recommended subject based on the third matrix.

In some embodiments, the first processing device 110, the second processing device 120, and the third processing device 130 may be the same processing device or may be different processing devices. A processing device may refer to any system having computing capabilities, and may include various computers such as servers and personal computers, as well as computing platforms consisting of multiple computers connected in a variety of configurations.

In some embodiments, the first processing device 110, the second processing device 120, and the third processing device 130 may include a Central Processing Unit (CPU), a Digital Signal Processor (DSP), a system on chip (SoC), a microcontroller unit (MCU), a computer, a user console, and the like, or any combination thereof. In some embodiments, the first processing device 110, the second processing device 120, and the third processing device 130 may comprise a single processing device or a group of processing devices. The processing device group may be centralized or distributed. In some embodiments, the first processing device 110, the second processing device 120, and the third processing device 130 may be local or remote. In some embodiments, the first processing device 110, the second processing device 120, and the third processing device 130 may be implemented on a cloud platform. For example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, or the like, or any combination thereof.

The user terminal 140 may enable user interaction (e.g., recommendation subject, recommendation object) with the intelligent recommendation system 100. For example, the recommended object may create a recommended object homepage in the intelligent recommendation system 100 through the user terminal 140, and upload presentation information and/or share contents on the recommended object homepage. For another example, the recommender may tag information and content of the intelligent recommendation system 100 via the user terminal 140. In some embodiments, user terminal 140 may include a mobile device 140-1, a tablet computer 140-2, a laptop computer 140-3, a desktop computer 140-4, other input and/or output enabled devices, and the like, or any combination thereof.

Network 150 may facilitate the exchange of information and/or data. In some embodiments, one or more components of the intelligent recommendation system 100 (e.g., the first processing device 110, the second processing device 120, the third processing device 130, the user terminal 140) may send information and/or data to other components of the intelligent recommendation system 100 over the network 150. For example, the first processing device 110 may acquire the number of times of execution of the recommended object, actual service information, importance scores, and the like from the user terminal 140 via the network 150. As another example, the second processing device 120 may obtain the importance scores, tagging behaviors, etc. of the recommended subjects from the user terminal 140 via the network 150. For another example, the third processing device 130 may acquire the first matrix and the second matrix from the first processing device 110 and the second processing device 120, respectively, via the network 150, and transmit the target recommended object to the user terminal 140 through the network 150.

In some embodiments, network 150 may include any one or more of a wired network or a wireless network. In some embodiments, network 150 may include a cable network, a fiber optic network, a telecommunications network, the internet, a Local Area Network (LAN), a Wide Area Network (WAN), a Wireless Local Area Network (WLAN), a Metropolitan Area Network (MAN), a Public Switched Telephone Network (PSTN), a bluetooth network, a ZigBee network, a Near Field Communication (NFC), an intra-device bus, an intra-device line, a cable connection, or the like, or any combination thereof. In some embodiments, the network connection between the components of the intelligent recommendation system 100 may be in one of the manners described above, or may be in a variety of manners. In some embodiments, the network may be a point-to-point, shared, centralized, etc. variety of topologies or a combination of topologies.

In some embodiments, the intelligent recommendation system 100 may also include a storage device (not shown). A storage device may be used to store data, instructions, and/or any other information. For example, the storage device may store a demand index, an evaluation index, a trigger event, and the like. In some embodiments, the storage device may include Random Access Memory (RAM), read Only Memory (ROM), mass storage, removable memory, volatile read-write memory, and the like, or any combination thereof. In some embodiments, the storage device may be integrated or included in one or more other components (e.g., the first processing device 110, the second processing device 120, the third processing device 130, the user terminal 140) of the intelligent recommendation system 100.

It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of the present description. Many variations and modifications will be apparent to those of ordinary skill in the art, given the benefit of this disclosure. The features, structures, methods, and other features of the exemplary embodiments described herein may be combined in various ways to obtain additional and/or alternative exemplary embodiments. However, such changes and modifications do not depart from the scope of the present specification.

FIG. 2 is an exemplary block diagram of an intelligent recommendation system, according to some embodiments of the present description. As shown in fig. 2, the intelligent recommendation system 200 may include a first matrix acquisition module 210, a second matrix creation module 220, a third matrix acquisition module 230, and a recommendation module 240.

The first matrix acquisition module 210 may be configured to acquire a first matrix including at least one evaluation index by the first processing device. In some embodiments, the at least one evaluation index may be determined based on the at least one evaluation floating point value. In some embodiments, the at least one evaluation floating point value may be determined based on the at least one recommended object and the at least one quantized tag type. In some embodiments, the quantization tags may include at least one of a construction body, a construction grade, a construction style, a construction object, a construction type, and a design capability. In some embodiments, the first matrix acquisition module 210 may perform one or more of the following: the first processing device obtains at least one quantized tag type; for each recommended object, the first processing device determines at least one evaluation floating point value corresponding to the at least one quantized tag type based on the total number of execution services and the number of execution services corresponding to the at least one quantized tag type. In some embodiments, the first matrix acquisition module 210 may perform one or more of the following: the first processing equipment determines at least one evaluation floating point value to be adjusted and at least one corresponding adjustment amplitude from at least one evaluation floating point value based on actual service information; the first processing device expands a range of corresponding at least one evaluation floating point value to be adjusted based on the at least one adjustment magnitude. In some embodiments, the actual business information may include at least one of recommended object activity, recommended object business completion information, and recommended subject feedback information. In some embodiments, the first matrix acquisition module 210 may perform one or more of the following: the first processing device determines at least one evaluation floating point value to be corrected and at least one event trigger value corresponding to the at least one evaluation floating point value based on the trigger event; the first processing equipment determines at least one event regression value corresponding to at least one floating point value to be corrected based on the time distance between the triggering time point of the triggering event and the current time point; the first processing device corrects the corresponding at least one to-be-corrected evaluation floating point value according to the at least one event trigger value and the at least one event regression value to obtain the corresponding at least one corrected evaluation floating point value. In some embodiments, the at least one modified evaluation floating point value may not exceed the range of the corresponding at least one evaluation floating point value to be adjusted. In some embodiments, the first matrix acquisition module 210 may perform one or more of the following: the first processing equipment acquires at least one first weight corresponding to at least one evaluation floating point value based on the importance of each quantization tag type to each recommended object; the first processing device determines a corresponding at least one evaluation index based on the at least one evaluation floating point value and the corresponding at least one first weight.

The second matrix creation module 220 may be configured to create a second matrix for the second processing device that includes at least one demand indicator. In some embodiments, the at least one demand indicator may be determined based on at least one demand floating point value. In some embodiments, the at least one demand floating point value may be determined based on the at least one recommended subject and the at least one quantized tag type. In some embodiments, the second matrix creation module 220 may perform one or more of the following: the second processing device obtains at least one required floating point value corresponding to at least one quantized tag type from at least one recommended subject; the second processing equipment obtains at least one second weight corresponding to the at least one required floating point value based on the preference degree of each recommending body for each quantized tag type; the second processing device determines at least one corresponding demand indicator according to the at least one demand floating point value and the corresponding at least one second weight to obtain a second matrix. In some embodiments, the second matrix creation module 220 may perform one or more of the following: the second processing device pushes at least one recommended object to obtain marking behaviors of the at least one recommended object by the at least one recommended object; the second processing equipment obtains at least one corresponding interest degree based on the marking behavior of each recommended object on at least one recommended object, and determines at least one interested recommended object based on the at least one interest degree; and for each recommended subject and each quantized label type, the second processing device acquires a requirement index corresponding to each recommended subject according to at least one evaluation index and at least one interestingness corresponding to at least one interested recommended object.

The third matrix obtaining module 230 may be configured to obtain a third matrix according to the first matrix and the second matrix by the third processing device.

The recommendation module 240 may be configured to determine, by the third processing device, the corresponding at least one target recommended object according to the target recommended subject based on the third matrix.

In some embodiments, the first matrix acquisition module 210, the second matrix creation module 220, the third matrix acquisition module 230, and the recommendation module 240 may be implemented on the same or different processing devices.

FIG. 3 is an exemplary flow chart of an intelligent recommendation method according to some embodiments of the present description.

In some embodiments, the first processing device 110, the second processing device 120, the third processing device 130, and/or the intelligent recommendation system may perform the process 300. As shown in fig. 3, the process 300 may include the following steps.

In step 310, the first processing device obtains a first matrix comprising at least one evaluation index. Specifically, step 310 may be performed by the first matrix acquisition module 210.

The evaluation index may be an index that evaluates a degree of recommendation of the ability of the recommended client. The recommended object may be a recommended object. By way of example only, in a construction application scenario, the recommended object may be a construction unit (e.g., a company, b company, c company, etc.), and the evaluation index may evaluate a degree of recommendation for a construction work capacity and/or design capacity of the construction unit. For example, the larger the value of the evaluation index, the more recommended the corresponding capability to the corresponding recommended object.

In some embodiments, the evaluation index may correspond to different quantized tag types. The detailed description of the quantization tags can be referred to as the related description of step 410, and will not be repeated here. For example, the evaluation index "design ability of American style" of the construction unit may correspond to the quantized tag type "American style". For another example, the evaluation index "quality of construction of foundation" of the construction unit may correspond to the quantization tag type "foundation".

The evaluation floating point value may be used to evaluate the size of the recommended guest capability. For example, the larger the construction work capacity of a construction unit, the larger the corresponding evaluation floating point value. By way of example only, the construction job capacities of the different construction grades of the first company are ordered as: "high end" = "low end" < "medium end", the corresponding evaluation floating point values may be 0.2=0.2 < 0.6, respectively.

In some embodiments, the at least one evaluation floating point value may be determined based on the at least one recommended object and the at least one quantized tag type. A detailed description of determining the at least one evaluation floating point value may be found in fig. 4 and its related description, and will not be repeated here.

The first weight may be used to evaluate the importance of the recommended guest capabilities to the recommended guest. For example, the more important a construction unit considers a certain construction job capability, the greater the first weight of the evaluation floating point value corresponding to the quantization tag type. In some embodiments, the sum of the first weights corresponding to different quantized tag types of the same quantized tag is equal to 1. For example only, the first weights corresponding to the construction grade types "high-end", "medium-end", and "low-end" of the first company are 0.5, 0.3, and 0.2, respectively.

In some embodiments, the first processing device may obtain, from the user terminal of the recommended object, an importance score corresponding to at least one quantized tag type of the recommended object, and then obtain, for each recommended object, at least one first weight corresponding to at least one evaluation floating point value based on the importance score of each quantized tag type of the same quantized tag. For example only, importance scores may be from 1 to 10 indicating that the importance is from low to high. For example, the first processing device may obtain importance scores of the first company construction grade types "high-end", "middle-end" and "low-end" from the user terminal of the first company as 10, 6 and 4, respectively, so as to determine that the corresponding first weights are 10/(10+6+4) =0.5, 6/(10+6+4) =0.3 and 4/(10+6+4) =0.2, respectively.

In some embodiments of the present disclosure, the determining the first weight based on the importance score provided by the recommended object may enable the first weight to embody a subjective intention of the recommended object to execute the service, thereby improving satisfaction of the recommended object with the recommendation result. For example, in the construction work of the first company, the evaluation floating point value corresponding to the construction grade "middle end" is the largest, that is, the construction work experience of the first company "middle end" is the most abundant, so the intelligent recommendation system may recommend the first company to more recommended subjects who need the construction grade of the "middle end" so that the construction work of the first company is more, however, the first company may be more prone to execute more construction works of the "high end", and more service problems of the "middle end" recommended to the first company may be broken based on the first weight. For another example, in the type of "construction quality" of the first company, the "micro cement" is a scarce process, the difficulty is large, so that recommended objects for which construction operations can be performed are small, the price is high, the market demand is small, so that recommended subjects having demands are small, and the number of times of the "micro cement" construction operations performed by the first company is correspondingly small, however, in order to recommend the first company to the recommended subjects having the process demands, the first company may set the corresponding first weight to be high.

In some embodiments, the at least one evaluation index may be determined based on the at least one evaluation floating point value. Specifically, the first processing device may determine the corresponding at least one evaluation indicator according to the at least one evaluation floating point value and the corresponding at least one first weight. For example only, the first processing device may take a product of the at least one evaluation floating point value and the corresponding at least one first weight as the corresponding evaluation index. For example, the evaluation indexes corresponding to "high end", "middle end", and "low end" of the a company may be 0.2×0.5=0.1, 0.6×0.3=0.18, and 0.2×0.2=0.04, respectively.

The first matrix may be used to evaluate at least one capability of at least one recommended object. In some embodiments, the dimension of the first matrix may be the number of recommended objects multiplied by the number of quantized tag types. For example, M recommended objects, N quantized tag types, the dimension of the first matrix is mxn.

In some embodiments of the present disclosure, an evaluation index is determined based on an evaluation floating point value and a first weight, so as to determine a first matrix, and the capability of the recommended object may be evaluated from objective experience and subjective willingness of the recommended object to execute a service, so that on one hand, the recommended object having the corresponding capability may be recommended to the recommended subject, and on the other hand, the recommended subject may be matched based on the willingness of the recommended object, and meanwhile, the exposure possibility of the rare process of the public may be improved.

At step 320, the second processing device establishes a second matrix including at least one demand index. In particular, step 320 may be performed by the second matrix creation module 220.

The requirement index may be an index for evaluating the degree of requirement of the recommending body for the business capability. The recommendation body may be a recommended object. For example only, in a construction application scenario, the recommended subject may be a customer (e.g., customer a, customer B, customer C, etc.) that needs a construction job service, and the demand index may evaluate the extent of demand for the customer for construction job capability and/or design capability. For example, the greater the value of the demand index, the higher the demand level of the corresponding recommended subject for the corresponding business capability. Similar to the evaluation index, in some embodiments, the demand index may correspond to a different quantized tag type. For example, the customer's demand index "American style design ability" and "construction quality of base dress" may correspond to the quantized tag types "American style" and "base dress", respectively. In some embodiments, the at least one demand indicator may be determined based on at least one demand floating point value.

The demand floating point value may be used to evaluate the recommended subject's importance to business capabilities. For example, the more important a customer considers to be in terms of a certain construction job's ability, the greater the required floating point value corresponding to the quantized tag type.

In some embodiments, the second processing device may obtain at least one required floating point value corresponding to the at least one quantized tag type from the at least one recommended subject. For example, the second processing device may obtain, from a user terminal of a certain recommendation entity, a importance score corresponding to at least one quantized tag type of the recommendation entity, and then obtain, based on the importance score corresponding to each quantized tag type of the same quantized tag by the recommendation entity, at least one required floating point value corresponding to the recommendation entity. For example only, a degree of importance score may be from 1 to 10 indicating that the recommended subject's degree of importance is from low to high. For example, the second processing device may obtain, from the user terminal of the client a, the importance scores of the client a for the construction grade types "high-end", "middle-end" and "low-end" as 1, 7, and 2, respectively, so that the required floating point values corresponding to the client a and "high-end", "middle-end" and "low-end" are 1/(2+7+1) =0.1, 7/(2+7+1) =0.7, and 2/(2+7+1) =0.2, respectively. Still further exemplary, the second processing device may obtain historical traffic demand data for a recommending entity from a user terminal of the recommending entity and determine the at least one demand floating point value based on the historical traffic demand data. For a detailed description of determining the demand floating point value based on the historical business demand data, reference may be made to fig. 4 for a description of determining the evaluation floating point value based on the number of times of execution business of the recommended object, which will not be repeated herein.

In some embodiments of the present disclosure, the requirement floating point value is determined based on the importance score provided by the recommending entity, so that the requirement floating point value can embody a preset subjective intention of the recommending entity on the requirement of the business capability, thereby improving satisfaction of the recommending entity on the recommending result.

The second weight may be used to evaluate a recommendation subject's preference for business capabilities. For example, the more preferred a business capability is by customer a, the greater the corresponding second weight. In some embodiments, the second processing device may obtain at least one second weight corresponding to the at least one required floating point value based on a tagging behavior of each of the recommended subjects for each of the quantized tag types.

The tagging behavior may be an interesting behavior of the recommending subject. For example, the act of marking the quantized tags by the recommending entity may include the recommending entity browsing, praying, collecting, sharing, etc. the content corresponding to the at least one quantized tag type at the user terminal for a long time. For another example, the act of tagging the quantized tags by the recommending entity may include the recommending entity jumping to a recommended guest conversation page that shares content based on content corresponding to at least one quantized tag type. The content corresponding to the quantized tag type may be display information and/or shared content of the recommended object in the intelligent recommendation system, and the detailed description of the display information and the shared content may refer to the related description of step 430, which is not repeated herein.

In some embodiments, the second processing device may count the number of times of tagging behavior of the recommendation body with respect to the at least one quantized tag type corresponding content based on the user terminal of the recommendation body, and determine the corresponding second weight based on the number of times of tagging behavior of each quantized tag type corresponding content. By way of example only, the second processing device may determine that the corresponding second weights are 15/(15+3+2) =0.75, 3/(15+3+2) =0.15, and 15/(15+3+2) =0.1, based on the user terminal of the client a, counting that the client a has collected the content corresponding to the "high end", "middle end", and "low end" of the construction level 15 times, 3 times, and 2 times, respectively.

In some embodiments, the second processing device may determine the corresponding at least one demand indicator based on the at least one demand floating point value and the corresponding at least one second weight to obtain the second matrix.

For example only, the second processing device may take as the corresponding demand indicator a product of the at least one demand floating point value and the corresponding at least one second weight. For example, the demand indicators corresponding to the "high-end", "middle-end", and "low-end" of the customer a may be 0.1×0.75=0.075, 0.7×0.15=0.108, and 0.2×0.1=0.02, respectively.

In some embodiments of the present description, the determining the second weight based on the tagging behavior of the recommending entity may mine the potential willingness of the recommending entity to the business capability requirement based on the second weight, thereby improving the satisfaction of the recommending entity with the recommending result. For example, the client a initially considers that the construction capacity of the "middle end" of the construction unit is more important based on objective factors such as price, the demand floating point values corresponding to the "high end" and the "middle end" are respectively 0.1 and 0.7, however, the client a prefers the "high end" service in the process of browsing the content of the intelligent recommendation system, the ratio of the "high end" demand index 0.075 and the "middle end" demand index 0.108 further acquired by the second service platform based on the second weight is respectively 0.075/(0.075+0.108+0.02) =0.37 and 0.108/(0.075+0.108+0.02) =0.53, so that the demand index of the "high end" service capacity preferred by the client a can be improved based on the second weight.

The second matrix may be used to evaluate a degree of demand for the at least one capability by the at least one recommending entity. In some embodiments, the dimension of the second matrix may be the number of recommended subjects multiplied by the number of quantized tag types. For example, K recommended objects, N quantized tag types, the dimension of the second matrix is kxn.

In some embodiments of the present disclosure, the requirement index is determined based on the requirement floating point value and the second weight of the recommendation body, so as to obtain the second matrix, and the explicit requirement and the potential requirement of the recommendation body can be simultaneously considered, so that the accuracy of the requirement index is improved.

It will be appreciated that the second processing device may not be able to obtain the required floating point value from the user terminal of the recommending entity. For example, the recommended subject has no preliminary preference for the importance of business capabilities. For another example, the recommendation entity may be adapted to provide a respective importance score for at least one quantized tag type. Thus, in some embodiments, the second processing device may determine the demand indicator based on the tagging behavior of the recommended subject to the recommended object.

In some embodiments, the second processing device may push the at least one recommended object to the at least one recommended subject to obtain a tagging behavior of the at least one recommended object by the at least one recommended subject. By way of example only, the client a does not provide a importance score corresponding to the quantized tag type, and the second processing device may then randomly push a preset number (e.g., 20) of construction units to the client a, and then obtain, from the user terminal of the client a, the client a's browsing, collecting, praying, and initiating a conversation, etc. of the marking actions of the preset number of construction units.

In some embodiments, the second processing device may obtain the corresponding at least one interest level based on the tagging behavior of each recommended subject to the at least one recommended object, and determine the at least one recommended object of interest based on the at least one interest level.

The interestingness may be an interestingness of the recommended subject with respect to the recommended object. In some embodiments, for each recommended subject, the second processing device may determine the degree of interest of the recommended subject for each recommended subject based on the recommended subject's tagging behavior for each recommended subject. By way of example only, when a recommending subject browses the home page of a recommending object, the interestingness of the corresponding recommending object may be increased by 0.1; when the time of browsing the homepage of the recommended object by the recommended subject exceeds 1 minute, the interestingness of the corresponding recommended object can be increased by 0.05; when the recommended subject collects the homepage of the recommended object, the interestingness of the corresponding recommended object can be increased by 0.2; when the recommending host initiates a dialogue to the recommending object, the interestingness of the corresponding recommending object can be increased by 0.3. For example, if customer a has collected the home page of the delta company and browsed for 2 minutes, the interest level of customer a in the delta company may be 0.1+0.05x2+0.2=0.4.

Further, the second processing device may determine the recommended object having the interestingness greater than the interestingness threshold as the interested object. For example, if the threshold of interest is 0.3, then the carrier may be determined to be a recommendation object of interest to customer A.

In some embodiments, for each recommended subject and each quantized tag type, the second processing device may obtain the requirement index corresponding to each recommended subject according to at least one evaluation index and at least one interestingness corresponding to at least one recommended object of interest. Specifically, for each recommended subject and each quantized label type, the maximum value and the minimum value in at least one evaluation index corresponding to at least one recommended object of interest may be respectively taken as the upper limit and the lower limit of the range of the requirement index corresponding to the recommended subject. For example, the recommended objects of interest of the client a include a company t, a company c and a company v, and evaluation indexes corresponding to the "high end" of the quantized tag types of the company t, the company c and the company v are respectively 0.5, 0.9 and 0.4, and then the upper limit and the lower limit of the requirement index range of the client a are respectively 0.9 and 0.4, that is, the requirement index range corresponding to the "high end" of the quantized tag types of the client a is 0.4-0.9.

In some embodiments, the difference between the upper limit and the lower limit of the range of the requirement index corresponding to the equivalent tag type exceeds the range threshold, and at least one recommended object may be randomly pushed to the recommended subject again to redetermine the range of the requirement index. For example, when the range threshold is 0.6 and the range of the requirement index corresponding to the "high end" of the quantized tag type of the client a is 0.2-0.9, the corresponding range of the requirement index can be redetermined.

Further, for each recommended subject and each quantized tag type, the second processing device may acquire a ratio of the interestingness and the evaluation index corresponding to each recommended object of interest, and use the ratio in the corresponding requirement index range as the corresponding requirement index. For example, for the client a and the quantized tag type "high end", the second processing device may obtain the ratio of the interestingness 0.4 of the interested recommended object "delta" to the interestingness 0.8 of the corresponding "high end" evaluation index "0.5" of "Ding Gongsi", and within the corresponding requirement index range 0.4-0.9, then 0.8 may be used as the requirement index corresponding to the client a and the quantized tag type "high end".

In some embodiments of the present disclosure, acquiring the requirement index of the recommendation subject based on the marking behavior of the recommendation subject on the recommendation subject may avoid inaccuracy of the requirement index caused by the deficiency of the requirement information of the recommendation subject.

In step 330, the third processing device obtains a third matrix according to the first matrix and the second matrix. Specifically, step 330 may be performed by the third matrix acquisition module 230.

The third matrix may be used to evaluate a degree of matching of the at least one recommended subject and the at least one recommended object. In some embodiments, the dimension of the third matrix may be the number of recommended objects multiplied by the number of recommended subjects. For example, M recommended objects and K recommended objects, the dimension of the third matrix is m×k.

In some embodiments, the third processing device may take, as the corresponding degree of matching, a product of an evaluation index corresponding to the same quantized tag type in the first matrix and a demand index corresponding to the second matrix. For example, the first matrix may be Q _M×N The second matrix may be represented by P _K×N Representation, then third matrix T _M×K =Q _M×N ·（P _K×N ） ^T 。

In step 340, the third processing device determines, based on the third matrix, at least one corresponding target recommended object according to the target recommended subject. Specifically, step 340 may be performed by recommendation module 240.

The target recommendation body may be a recommendation body having recommendation requirements. In some embodiments, the recommendation module 240 may determine the target recommendation body based on the requesting behavior of the recommendation body. In some embodiments, the requesting act may include the recommender actively initiating a matching request. For example, client a may send a match request to the third processing device by clicking on the "match" button of the user terminal. In some embodiments, the requesting action may include the recommender searching, browsing, and/or publishing a recommendation object need at the intelligent recommendation system.

The target recommended object may be a recommended object that matches the target recommended subject. In some embodiments, the evaluation index of the target recommended object may be matched with the demand index of the target recommended subject. Specifically, for each target recommended subject, the recommendation module 240 may determine at least one matching degree of at least one recommended object based on the third matrix, and rank the at least one recommended object based on the at least one matching degree, and determine the recommended object with the largest matching degree as the target recommended object. For example based on a third matrix T _M×K The K recommended objects of the client A can be ranked according to the K matching degrees of the 1 st column (namely, the column corresponding to the client A), and the recommended object company B corresponding to the maximum value in the K matching degrees is determined as the target recommended object of the client A.

In some embodiments, the recommendation module 240 may also recommend a target recommendation subject to the target recommendation object. Specifically, the recommendation module 240 may determine the target recommended object based on the request behavior of the recommended object, and then determine the corresponding target recommended subject based on the third matrix.

It should be noted that the above description of flow 300 is provided for illustrative purposes only and is not intended to limit the scope of the present description. Various changes and modifications may be made by one of ordinary skill in the art in light of the description herein. However, such changes and modifications do not depart from the scope of the present specification. In some embodiments, the process 300 may include one or more additional operations, or one or more of the operations described above may be omitted.

In some embodiments of the present disclosure, the evaluation index and the demand index of the recommended object and the recommended subject are respectively linearized based on the quantization tag type, so that the matching accuracy of the target recommended subject and the target recommended object, which are matched based on the evaluation index and the demand index, is higher.

FIG. 4 is an exemplary flow chart for determining an evaluation floating point value according to some embodiments of the present description.

In some embodiments, the first processing device 110, the intelligent recommendation system 200, and/or the first matrix acquisition module 210 may perform the flow 400. In some embodiments, step 310 in fig. 3 may be implemented by flow 400. As shown in fig. 4, the flow 400 may include the following steps.

At step 410, the first processing device obtains at least one quantized tag type.

The quantization tag may be an evaluation dimension of recommended customer capability, for example, an evaluation dimension of construction job capability of a construction unit. Illustratively, the quantization tag of the construction unit may include at least one of a construction subject, a construction grade, a design style, a construction object, a construction type, and a construction quality. The construction subject can be a construction requirement party, the construction object can be a constructed building and/or a site, and the construction type can be the operation type of construction engineering.

In some embodiments, each quantization label may be determined based on a last level type of each evaluation dimension. For example, taking a construction object as an example, the types of the construction object may include storefront, office, home decoration and public area, wherein the types of the home decoration may be further divided into private home, duplex, large flat, LOFT, villa, and the type of the quantization tag of the "construction object" may include storefront, office, private home, duplex, large flat, LOFT, villa and public area.

Illustratively, the types of construction subjects may include business units, public institutions, individual families, individual merchants, property, communities, etc., the types of construction grades may include high-end, medium-end, low-end, etc., the types of design styles may include chinese, japanese, northern europe, french, american, etc., the types of construction may include foundation finishing, reconstruction engineering, reinforcement engineering, maintenance engineering, sporadic engineering, micro-cement construction, etc., and the quality of construction may include base quality, micro-cement construction quality, etc.

In some embodiments, the quantized tag type may be preset based on industry standards. For example, a quantized tag type of a construction application scenario may be determined based on construction acceptance criteria. In some embodiments, the quantified tag type may be confirmed by an administrator with rights after being uploaded by the recommended object. For example, the first company may upload its adept scarce construction process "micro-cement" as a type of quantized tag to the intelligent recommendation system 200, and the authorized manager may confirm the "micro-cement" as a type of construction quality.

In step 420, for each recommended object, the first processing device determines at least one evaluation floating point value corresponding to the at least one quantized tag type based on the total number of execution services and the number of execution services corresponding to the at least one quantized tag type.

The total number of execution of the service may be a total number of execution of the service by the recommended object. For example, the total number of times the construction unit performs the construction work. By way of example only, the total number of construction operations of the first company is 100.

The number of times of executing the service corresponding to the at least one quantized tag type may refer to the number of times of executing the service belonging to the quantized tag type in the recommended object executing the service. It can be understood that the sum of the number of times of executing the service corresponding to different quantized tag types of the same quantized tag is equal to the total number of times of executing the service. For example, fig. 5 is an exemplary flowchart for determining and evaluating floating point values according to some embodiments of the present disclosure, and as shown in fig. 5, in the construction work of the company a, the number of times of construction belonging to the high-end, middle-end, and low-end is 20, 60, and 20, respectively, for a total of 100.

From the foregoing, the evaluation floating point value may be used to evaluate the size of the recommended guest capability. In some embodiments, for each recommended object, the evaluation floating point value corresponding to each quantized tag type may be a ratio of the number of execution traffic times corresponding to the quantized tag type to the total number of execution traffic times. For example, as shown in fig. 5, type of "construction grade" by a company a: the evaluation floating point values corresponding to the high end, the terminal and the low end are respectively 20/100=0.2, 60/100=0.6 and 20/100=0.2.

In some embodiments of the present disclosure, based on experience of a recommended object executing a service, capabilities of the recommended object are converted into evaluation floating point values that can be linearly quantized from multiple dimensions through different quantization tags, and a complex recommended object evaluation system is converted into structured data of interest of the service, so that accuracy of subsequent evaluation of the recommended object is improved.

The first processing device determines at least one evaluation floating point value to be adjusted and corresponding at least one adjustment magnitude from the at least one evaluation floating point value based on the actual traffic information, step 430.

The actual service information may be information reflecting the actual situation when the recommended object performs the service. In some embodiments, the actual business information may include at least one of recommended object activity, recommended object business completion information, and recommended subject feedback information.

The recommended object liveness may be the liveness of the recommended object in the intelligent recommendation system. In some embodiments, the recommended object liveness may include, but is not limited to, a speed of response of the recommended object to the recommended subject initiating a conversation and/or a dispatch, a number of shares of the recommended object's excellent design cases, a number of shares of the recommended object's excellent process cases, and so on.

In some embodiments, the first processing device may determine the activity of the recommended object based on the user terminal of the recommended object counting the behavior (e.g. response, sharing, etc.) of the recommended object. For example, the liveness of the recommended objects with a response time ranking of 20% or a sharing time ranking of 20% may be determined to be high.

In some embodiments of the present disclosure, the intelligent recommendation system may be caused to actively use the intelligent recommendation system based on the activity of the recommendation object, so that the intelligent recommendation system generates more content to provide to the recommendation subject, which improves the use experience of the recommendation subject on the one hand, and improves the recommendation accuracy based on the interest degree of the recommendation subject in the content on the other hand.

The recommended object service completion information may be presentation information of the recommended object completed service. By way of example only, presentation information recommending that the object have completed the business may include a scarcity process and/or a uniquely designed construction completion image, text presentation, etc. For example, construction of "micro-cement" completes video and text presentation. For another example, the construction of the "design of a special house type" completes the 3D image.

In some embodiments of the present disclosure, by presenting the recommended object with a completed service of a scarce service (e.g., "micro-cement" or "design of a special-shaped house"), the matching efficiency of the recommended object and the recommended object of the scarce service can be improved, so that the matching success rate of the recommended object and the recommended object of the scarce service is improved.

The recommendation subject feedback information may be feedback information of the history recommendation subject to the recommendation subject. In some embodiments, the recommended subject feedback information may include, but is not limited to, design issue feedback (e.g., not delivering design manuscripts on schedule, design manuscripts not achieving desired effects, design manuscript satisfaction, etc.), construction issue feedback (e.g., not completing construction on schedule, construction quality not reaching standards, not constructing according to design manuscripts, construction satisfaction, etc.), after-market aggressiveness feedback (e.g., after-market response time, after-market service attitudes, etc.), and price feedback (e.g., price added, mixed charge during construction, etc.), etc. In some embodiments, the first processing device may receive feedback information of the history recommendation subject on the recommendation object from a user terminal of the history recommendation subject.

In some embodiments of the present disclosure, the problem that the actual service execution situation (for example, the site construction situation) and the recommended object evaluation floating point value are not matched may be solved by the feedback information of the recommended subject, so that the accuracy of evaluating the capability of the recommended object is improved based on the angle of the recommended subject.

The evaluation floating point value to be adjusted may be an evaluation floating point value of a correction range to be extended. It can be understood that the actual situation and the evaluation floating point value when the recommended object executes the service cannot be completely matched due to the relationship between the recommended subject and the recommended object, the difference of the overall performance of the service of the recommended object, the change of the industry environment and the like. For example, although the evaluation floating point value 0.6 corresponding to the quantization tag type "middle end" of the company a is higher than the evaluation floating point value 0.5 corresponding to the company b, the total number of construction operations of the company a is 100, the number of construction operations of the company a "middle end" is 60, the total number of construction operations of the company b is 200, the number of construction operations of the company b "middle end" is 100, and in fact, the construction operation experience of the company b "middle end" is more abundant than that of the company a. For another example, the evaluation floating point value corresponding to the quantized tag type "middle end" of the third company is 0.6 higher than the evaluation floating point value corresponding to the fourth company by 0.5, the total number of construction operations of the fourth company and the third company is 200, the number of construction operations of the fourth company is 100, the number of construction operations of the third company is 120, and the construction operation experience of the third company is richer than that of the fourth company, however, in the process of performing the construction operation of the third company, the experience of the third company is often reduced due to contradiction between the super budget and the recommended main body. Therefore, when the actual situation of the recommended object in the field changes more when the service is executed, the range of the corrected evaluation floating point value is correspondingly larger. The detailed description of the correction evaluation floating point value can be referred to steps 450 to 470, which are not repeated here.

In some embodiments, the first processing device may determine the evaluation floating point value to be adjusted based on the recommended object liveness. For example only, the first processing device may determine the corresponding evaluation floating point value to be adjusted based on the type of quantization tag corresponding to the excellent design case or the excellent process case shared by the recommended objects. For example, based on the micro-cement process case shared by the first company, it may be determined that the evaluation floating point value corresponding to the first company and the quantized tag type "micro-cement" is 0.01 as the evaluation floating point value to be adjusted.

In some embodiments, the first processing device may determine the evaluation floating point value to be adjusted based on the recommended object service completion information. For example only, the first processing device may determine the corresponding evaluation floating point value to be adjusted based on the quantization tag type corresponding to the completed service exhibited by the recommended object. For example, based on the construction completion information of the micro-cement process displayed by the first company, it is determined that the evaluation floating point value corresponding to the first company and the quantized tag type "micro-cement" is 0.01 as the evaluation floating point value to be adjusted.

In some embodiments, the first processing device may determine the evaluation floating point value to be adjusted based on the historical recommended subject feedback information. For example only, the first processing device may determine the corresponding evaluation floating point value to be adjusted based on the quantization tag type and the recommended object corresponding to the historical recommended subject feedback information. For example, based on the historical recommended subject feedback information "the quantization tag types" Chinese "corresponding to the" a company did not deliver the design draft on schedule, "…," American ", and the corresponding recommended object" a company, "the evaluation floating point values 0.3, …, 0.1 corresponding to" Chinese ", …," American "and" a company "are determined as the evaluation floating point values to be adjusted.

The adjustment amplitude may be a variable amplitude of the evaluation floating point value to be adjusted at the time of matching. The adjustment amplitude can reflect the influence of the actual service information on the change of the evaluation floating point value. For example, the larger the influence, the larger the adjustment amplitude.

In some embodiments, the first processing device may determine the corresponding adjustment amplitude based on the recommended object liveness. For example only, when the recommended object activity is high, the upper limit of the adjustment amplitude may be increased by 5%; when the recommended object activity is middle, the adjustment amplitude is unchanged; when the recommended object activity is low, the lower limit of the adjustment range is reduced by 5%. For example, if the recommended object activity of the first company is high, the upper limit of the adjustment range of the to-be-adjusted evaluation floating point value corresponding to the first company may be increased by 5%.

In some embodiments, the first processing device may determine the corresponding adjustment amplitude based on the recommended object service completion information. For example only, the first processing device may determine the corresponding adjustment amplitude based on the quality, number (e.g., number of pictures, video length, number of words introduced, etc.) of the recommended object completed service presentation information and the presentation rule. For example, the upper limit of the adjustment range is increased by 5% when the length of the video is 5-10 min based on the length ' 10min ' and the display rule ' of the construction completion video of the micro cement of the A company, and the upper limit of the adjustment range of the floating point value to be evaluated corresponding to the A company is increased by 5%.

In some embodiments, the first processing device may determine the adjustment amplitude based on historical recommendation subject feedback information. For example only, the first processing device may determine the corresponding adjustment amplitude based on the design draft delivery expiration time and the feedback rule fed back by the historic recommendation body. For example, based on the historical recommendation body feedback information of 5 days of out-of-date time of the design draft which is not delivered on schedule by the first company and the feedback rule, when the out-of-date time of the design draft is 3-5 days, the lower limit of the adjustment range is reduced by 2%, and the lower limit of the adjustment range of the to-be-adjusted evaluation floating point value corresponding to the first company is reduced by 2%.

In step 440, the first processing device expands a range of the corresponding at least one evaluation floating point value to be adjusted based on the at least one adjustment magnitude.

Specifically, for each evaluation floating point value to be adjusted, the first processing apparatus may add an upper limit of the corresponding adjustment range and subtract a lower limit of the adjustment range, respectively, to thereby obtain an extended range of the evaluation floating point value to be adjusted. Continuing the above example, taking the to-be-adjusted evaluation floating point values corresponding to the "a company" and the "micro cement" as an example, the first processing device determines that the upper limit of the corresponding adjustment range can be increased by 5% +5% =10% based on the actual service information, and the lower limit is unchanged, and the corresponding to-be-adjusted evaluation floating point value range is 0.01-0.011.

In some embodiments, the first processing device may set a minimum threshold and a maximum threshold for the evaluation floating point value to be adjusted. For example, the minimum threshold is 0 and the maximum threshold is 1. In some embodiments, when the revised evaluation float value is less than the minimum threshold value, then taking the minimum threshold value as the revised evaluation float value; and when the correction evaluation floating point value is larger than the maximum threshold value, taking the maximum threshold value as the correction evaluation floating point value.

In some embodiments of the present disclosure, the range of the evaluation floating point value to be adjusted is extended based on the actual service information, so that the problem that the actual situation when the recommended object executes the service and the evaluation floating point value cannot be completely matched can be solved, errors caused by the relation between the recommended objects, the difference in the performance of the recommended object service, the change in the industry environment, and the like are balanced, and meanwhile, the recommended object can be stimulated through a variable extension range to improve the service capability corresponding to the quantization tag and the aggressiveness of executing the service.

At step 450, the first processing device determines at least one evaluation floating point value to be modified and a corresponding at least one event trigger value from the at least one evaluation floating point value based on the trigger event.

The triggering event may be an event that a third party produces an objective rating (including positive and negative ratings) for the recommended guest's ability. For example, the triggering event may include, but is not limited to, the recommended object obtaining an industry prize (i.e., an industry expert's positive evaluation of the recommended object), the recommended object being subject to administrative penalties (i.e., an administrative department's negative evaluation of the recommended object), the recommended object obtaining consumer approval (i.e., a consumer group's positive evaluation of the recommended object), and the like.

In some embodiments, the first processing device may obtain the trigger event through an input of the user terminal. By way of example only, the administrator may enter the trigger event through the user terminal such that the first processing device may receive the trigger event sent by the user terminal through the network. In some embodiments, the first processing device may also obtain the trigger event periodically through web crawling, information retrieval, and other interfaces, which is not limited in this embodiment.

The evaluation floating point value to be corrected is an evaluation floating point value that can be further corrected. It can be understood that the evaluation floating point value is determined based on experience of the recommended object executing the service, and the evaluation floating point value to be corrected can be further evaluated based on the evaluation of the recommended object by a third party through a trigger event, so that the dimension of evaluation is increased, and the accuracy of evaluation is improved.

In some embodiments, the first processing device may determine, based on the trigger event, the related recommended object and the ability of the related recommended object to be evaluated by the third party, determine the related quantized tag type based on the evaluated ability, and then determine the evaluation floating point value corresponding to the related recommended object and the related quantized tag type as the evaluation floating point value to be corrected.

By way of example only, the triggering event includes a first company obtaining administrative penalties for unqualified quality of construction by the market administration at 1 month 6 of 2022, then the associated recommended object may be determined to be the first company and the ability of the associated recommended object to be evaluated by a third party as the construction ability, thereby determining that the associated quantized tag type includes a construction object "storefront", … "regular house", a construction grade "high-end", "medium-end", "low-end", a construction quality "base", … "micro-cement", etc., and then determining the evaluation floating point values corresponding to the first company and the "storefront", … "regular house", "high-end", "medium-end", "low-end", "base", … "micro-cement" as the evaluation floating point values to be corrected.

As yet another example, the triggering event includes obtaining a "decade wind design jackpot" by 2022,

month

6, 1, company b, company c, etc., then the ability of the related recommended object to be evaluated by the third party may be determined as a chinese style design ability, thereby determining that the related quantized tag type is "chinese", and then the evaluation floating point values corresponding to company b, company c, and "chinese" are determined as the evaluation floating point values to be corrected.

As yet another example, as shown in fig. 5, the trigger event includes that the first company obtains the "best design prize" on 1 st 2022, then the related recommended object is determined to be the first company and the ability of the related recommended object to be evaluated by the third party is determined to be the design ability, so as to determine the related quantized tag type to include "chinese" and … "america" in design style, and then the evaluation floating point values corresponding to the first company and "chinese" and … "america" are determined to be the evaluation floating point values to be corrected.

The event trigger value may represent the magnitude of the impact of the trigger event on the evaluation floating point value to be modified. For example, the greater the impact of a trigger event on the evaluation floating point value to be modified, the greater the absolute value of the event trigger value. And when the triggering event is a positive evaluation event of the recommended guest ability by a third party, the event triggering value is positive, and otherwise, the event triggering value is negative.

In some embodiments, the first processing device may determine an event trigger value corresponding to the trigger event based on a preset rule. The preset rule may include event trigger values corresponding to different trigger events. In some embodiments, the preset rules may be preset by an administrator with rights. In some embodiments, the preset rules may be initiated by the recommended objects and validated by an administrator with rights. For example, the recommended object may transmit the candidate rule to the first processing device through the user terminal, and the first processing device may transmit the candidate rule to the user terminal of the administrator and receive whether to determine the candidate rule as the preset rule from the user terminal of the administrator.

For example only, the preset rules may include: when the triggering event is administrative punishment of 1-2 ten thousand yuan, the corresponding event triggering value is-0.01; when the administrative penalty is 'penalty of 2-5 ten thousand yuan', the corresponding event trigger value is-0.02; when the administrative penalty is ' penalty more than 5 ten thousand yuan ' or ' stop operation is complete, the corresponding event trigger value is-0.03; when the administrative penalty is "revoked business license" or "revoked related qualification", the corresponding event trigger value is equal to the current evaluation floating point value. For example, if the administrative penalty of the first company for obtaining the unqualified construction quality of the market supervision bureau is "fine 2-5 ten thousand yuan", the event trigger value corresponding to the first company and the "storefront", … "private", "high-end", "medium-end", "low-end", "base-load" and … "micro cement" is-0.02.

In step 460, the first processing device determines at least one event regression value corresponding to the at least one floating point value to be modified based on the time distance between the trigger time point of the trigger event and the current time point.

The trigger time point of the trigger event may be the generation time or the validation time of the trigger event. Continuing with the foregoing example, the trigger time point of the trigger time "the first company obtains the administrative penalty that the quality of the construction of the market management agency is not acceptable" may be 2022 years, 6 months, 1 day of the release time of the administrative penalty, and the current time point is 2022 years, 7 months, 1 day, and the time distance between the trigger time point of the trigger event "the first company obtains the administrative penalty that the quality of the construction of the market management agency is not acceptable" and the current time point is 30 days.

The impact decay rate may represent the decay rate at which the trigger event affects the floating point value to be modified. In some embodiments, the first processing device may determine the corresponding impact decay rate based on the occurrence period of the trigger event. For example only, the corresponding impact decay rate may be determined based on a ratio of an event trigger value of the trigger event to an occurrence period (e.g., a prize evaluation period). For example, the event trigger value of "Ten-good wind design jackpot" is 0.5 and the review period is once per year, then the corresponding impact decay rate may be 0.5/365. For another example, if the time trigger value of "best design prize" is 1 and the review period is one half year, the corresponding impact decay rate may be 1/182. In some embodiments, the first processing device may determine a corresponding impact elimination time based on an impact range of the trigger event (e.g., an administrative level of the administrative penalty subject), and determine a corresponding decay rate based on the corresponding event trigger value and the impact elimination time. For example, based on the trigger event "administrative penalty for unqualified quality of market administration" market administration being market level, the corresponding impact removal time can be determined to be 100 days, and based on the event trigger value of-0.02 and the impact removal time of 100 days, the corresponding impact decay rate can be determined to be-0.0002. In some embodiments, the rate of attenuation may be affected unevenly, and the present embodiment is not limited. For example, the rate of decay may be affected rapidly and then uniformly. As another example, influencing the decay rate may be slower and slower.

The event regression value may represent the extent to which the triggering event affects the decay of the floating point value to be modified. For example, the greater the absolute value of the event regression value, the greater the degree of impact attenuation. In some embodiments, the first processing device may determine an event regression value corresponding to the trigger event based on the time distance corresponding to the trigger event and the impact decay rate. For example only, the temporal regression value may be determined based on the product of the temporal distance and the impact decay rate. For example, based on the time distance of 30 days between the trigger time point at which the first company obtains the administrative penalty that the quality of the construction of the market supervision bureau is not acceptable and the current time point, and the influence attenuation speed-0.0002, it is determined that the event regression value corresponding to the first company and the "storefront", … "private", "high-end", "medium-end", "low-end", "base-load", … "micro cement" is-0.006.

In step 470, the first processing device corrects the corresponding at least one to-be-corrected evaluation floating point value according to the at least one event trigger value and the at least one event regression value, so as to obtain the corresponding at least one corrected evaluation floating point value.

The revised evaluation float may be an evaluation float corrected based on the third party objective evaluation. For each floating point value to be modified, the first processing device may add the corresponding event trigger value to the floating point value to be modified and subtract the corresponding event regression value to obtain the corresponding modified floating point value. Continuing with the example above, the revised evaluation floating point value for company a and "storefront" is 0.1+ (-0.02) - (-0.006) =0.086. Further, the first processing device determines that the revised evaluation floating point value corresponding to the first company and "Chinese" is 0.3+ (-0.02) - (-0.006) +1-1/182×181=0.086+1-181/182=0.081 based on the "best design prize" event trigger value 1 and the influence attenuation speed 1/182 obtained by the first company on 1 month 1 day 2022.

Similarly, the first processing device may determine that the correction evaluation floating point values corresponding to the company b, the company c, and the "chinese" are 0.5+0.5- (0.5/365×30) =0.96 and 0.6+0.5- (0.5/365×30) =1.06, respectively, based on the "decade majora design event trigger value of 0.5 and the influence decay rate of 0.5/365 on the 1 st day of 2022, 6 th month.

In some embodiments, the at least one modified evaluation floating point value may not exceed the range of the corresponding at least one evaluation floating point value to be adjusted. Specifically, when the modified evaluation floating point value is smaller than the lower limit of the range of the to-be-adjusted evaluation floating point value, taking the lower limit as the modified evaluation floating point value; and when the correction evaluation floating point value is larger than the upper limit of the range of the to-be-adjusted evaluation floating point value, taking the upper limit of the range as the correction evaluation floating point value. For example, if the modified evaluation floating point value corresponding to the "micro cement" of the company a is 0.01+ (-0.02) - (-0.006) = -0.004 and the range of the corresponding to-be-adjusted evaluation floating point value is 0.01-0.011, the lower limit 0.01 of the range can be used as the corresponding modified evaluation floating point value.

In some embodiments of the present disclosure, the evaluation floating point value may be adjusted in conjunction with industry third party evaluation by modifying the evaluation floating point value by a trigger event, thereby improving the authenticity of the recommended object evaluation index.

Possible benefits of embodiments of the present description include, but are not limited to: (1) Based on the quantization label type, linearizing the evaluation index and the demand index of the recommended object and the recommended subject respectively, so that the matching accuracy of the target recommended subject and the target recommended object matched based on the evaluation index and the demand index is higher; (2) Based on experience of recommended objects in executing service, the capability of the recommended objects is converted into evaluation floating point values which can be linearly quantized from multiple dimensions through different quantization tags, a complex recommended object evaluation system is converted into structured data of service attention, and the subsequent evaluation accuracy of the recommended objects is improved; (4) The range of the evaluation floating point value to be adjusted is expanded based on the actual service information, so that the problem that the actual situation and the evaluation floating point value cannot be completely matched when a recommended object executes the service can be solved, errors caused by the relation between a recommended subject and the recommended object, different performance of the recommended object service, environmental change of the industry and the like are balanced, meanwhile, the recommended object is stimulated through the variable expansion range to improve the service capacity corresponding to the quantized tag and the enthusiasm of the execution service, and meanwhile, the evaluation floating point value is corrected through a trigger event, and the evaluation floating point value can be adjusted by combining with the evaluation of a third party of the industry, so that the authenticity and the practicability of the evaluation index of the recommended object are improved; (5) And determining an evaluation floating point value based on the evaluation floating point value and the first weight, and evaluating the capability of the recommended object from objective experience and subjective willingness of the recommended object to execute the service, wherein on one hand, the recommended object with the corresponding capability can be recommended to the recommended object, on the other hand, the recommended object can be matched based on the willingness of the recommended object, and meanwhile, the exposure possibility of the rare process of the masses is improved.

While the basic concepts have been described above, it will be apparent to those skilled in the art that the foregoing detailed disclosure is by way of example only and is not intended to be limiting. Although not explicitly described herein, various modifications, improvements, and adaptations to the present disclosure may occur to one skilled in the art. Such modifications, improvements, and modifications are intended to be suggested within this specification, and therefore, such modifications, improvements, and modifications are intended to be included within the spirit and scope of the exemplary embodiments of the present invention.

Meanwhile, the specification uses specific words to describe the embodiments of the specification. Reference to "one embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the present description. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the present description may be combined as suitable.

Furthermore, the order in which the elements and sequences are processed, the use of numerical letters, or other designations in the description are not intended to limit the order in which the processes and methods of the description are performed unless explicitly recited in the claims. While certain presently useful inventive embodiments have been discussed in the foregoing disclosure, by way of various examples, it is to be understood that such details are merely illustrative and that the appended claims are not limited to the disclosed embodiments, but, on the contrary, are intended to cover all modifications and equivalent arrangements included within the spirit and scope of the embodiments of the present disclosure. For example, while the system components described above may be implemented by hardware devices, they may also be implemented solely by software solutions, such as installing the described system on an existing processing device or mobile device.

Likewise, it should be noted that in order to simplify the presentation disclosed in this specification and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof. This method of disclosure, however, is not intended to imply that more features than are presented in the claims are required for the present description. Indeed, less than all of the features of a single embodiment disclosed above.

In some embodiments, numbers describing the components, number of attributes are used, it being understood that such numbers being used in the description of embodiments are modified in some examples by the modifier "about," approximately, "or" substantially. Unless otherwise indicated, "about," "approximately," or "substantially" indicate that the number allows for a 20% variation. Accordingly, in some embodiments, numerical parameters set forth in the specification and claims are approximations that may vary depending upon the desired properties sought to be obtained by the individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and employ a method for preserving the general number of digits. Although the numerical ranges and parameters set forth herein are approximations that may be employed in some embodiments to confirm the breadth of the range, in particular embodiments, the setting of such numerical values is as precise as possible.

Each patent, patent application publication, and other material, such as articles, books, specifications, publications, documents, etc., referred to in this specification is incorporated herein by reference in its entirety. Except for application history documents that are inconsistent or conflicting with the content of this specification, documents that are currently or later attached to this specification in which the broadest scope of the claims to this specification is limited are also. It is noted that, if the description, definition, and/or use of a term in an attached material in this specification does not conform to or conflict with what is described in this specification, the description, definition, and/or use of the term in this specification controls.

Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments of this specification. Other variations are possible within the scope of this description. Thus, by way of example, and not limitation, alternative configurations of embodiments of the present specification may be considered as consistent with the teachings of the present specification. Accordingly, the embodiments of the present specification are not limited to only the embodiments explicitly described and depicted in the present specification.

Claims

1. An intelligent recommendation method, characterized in that the method comprises the following steps:

The first processing device obtains a first matrix comprising at least one evaluation index, the evaluation index being an index for evaluating a degree of recommendation of a recommended guest capability, the at least one evaluation index being determined based on at least one evaluation floating point value for evaluating a size of the recommended guest capability and a corresponding at least one first weight, the at least one evaluation floating point value being determined based on at least one recommended guest and at least one quantized tag type, the first weight being obtained for each degree of importance of the recommended guest based on each of the quantized tag types, the degree of importance of each recommended guest being provided by the recommended guest;

the second processing device establishes a second matrix comprising at least one demand indicator, the demand indicator being an indicator of the recommended subject's degree of demand for business capability, the at least one demand indicator being determined based on at least one demand floating point value for evaluating the recommended subject's degree of importance for business capability, the at least one demand floating point value being determined based on at least one recommended subject and the at least one quantized tag type;

The third processing equipment obtains a third matrix according to the product of the first matrix and the second matrix;

the third processing device determines at least one corresponding target recommended object according to the target recommended subject based on the third matrix.

2. The intelligent recommendation method of claim 1, wherein the at least one evaluation floating point value is determined based on at least one recommended object and at least one quantized tag type, comprising:

the first processing device obtaining the at least one quantized tag type;

for each recommended object, the first processing device determines at least one evaluation floating point value corresponding to the at least one quantized tag type based on a total number of execution services and a number of execution services corresponding to the at least one quantized tag type.

3. The intelligent recommendation method of claim 2, wherein the at least one evaluation floating point value is determined based on at least one recommended object and at least one quantized tag type, further comprising:

the first processing device determines at least one evaluation floating point value to be adjusted and at least one corresponding adjustment amplitude from the at least one evaluation floating point value based on actual service information;

The first processing device expands a range of the corresponding at least one evaluation floating point value to be adjusted based on the at least one adjustment magnitude.

4. The intelligent recommendation method of claim 3, wherein the at least one evaluation floating point value is determined based on at least one recommended object and at least one quantized tag type, further comprising:

the first processing device determines at least one evaluation floating point value to be modified and at least one event trigger value corresponding to the at least one evaluation floating point value based on a trigger event;

the first processing device determines at least one event regression value corresponding to the at least one to-be-corrected evaluation floating point value based on the time distance between the trigger time point of the trigger event and the current time point;

the first processing device corrects the corresponding at least one to-be-corrected evaluation floating point value according to the at least one event trigger value and the at least one event regression value to obtain the corresponding at least one corrected evaluation floating point value, wherein the at least one corrected evaluation floating point value does not exceed the range of the corresponding at least one to-be-adjusted evaluation floating point value.

5. The intelligent recommendation method of claim 1, wherein the second processing device establishes a second matrix comprising at least one demand indicator, comprising:

The second processing device obtains at least one required floating point value corresponding to the at least one quantized tag type from the at least one recommended subject;

the second processing device obtains at least one second weight corresponding to the at least one required floating point value based on the preference degree of each recommending body for each quantized tag type;

the second processing device determines the corresponding at least one demand indicator according to the at least one demand floating point value and the corresponding at least one second weight to obtain the second matrix.

6. The intelligent recommendation method of claim 1, wherein the second processing device establishes a second matrix comprising at least one demand indicator, comprising:

the second processing device randomly pushes at least one recommended object to the at least one recommended object to acquire the marking behavior of the at least one recommended object by the at least one recommended object;

the second processing device obtains at least one corresponding interest degree based on the marking behavior of each recommended subject on the at least one recommended object, and determines at least one recommended object of interest based on the at least one interest degree;

And for each recommended subject and each quantized label type, the second processing device acquires a requirement index corresponding to each recommended subject according to at least one evaluation index corresponding to the at least one interested recommended object and the at least one interest degree.

7. The intelligent recommendation method of claim 1, wherein the quantization tags include at least one of construction subject, construction grade, construction style, construction object, construction type, and design capability.

8. An intelligent recommendation system, the system comprising:

a first matrix acquisition module configured to acquire a first matrix including at least one evaluation index, the evaluation index being an index that evaluates a degree of recommendation of a recommended guest capability, the at least one evaluation index being determined based on at least one evaluation floating point value for evaluating a size of the recommended guest capability and corresponding at least one first weight, the at least one evaluation floating point value being determined based on at least one recommended guest and at least one quantized tag type, the first weight being acquired for importance of each recommended guest based on each of the quantized tag types, the importance of each recommended guest being provided by the recommended guest;

A second matrix building module, configured to build a second matrix including at least one requirement index, where the requirement index is an index for evaluating a requirement level of a recommended subject for a service capability, and the at least one requirement index is determined based on at least one requirement floating point value, where the requirement floating point value is used to evaluate a importance level of the recommended subject for the service capability, and the at least one requirement floating point value is determined based on at least one recommended subject and the at least one quantization tag type;

a third matrix acquisition module, configured to acquire a third matrix according to a product of the first matrix and the second matrix by a third processing device;

and the recommendation module is used for determining at least one corresponding target recommended object according to the target recommended subject based on the third matrix by the third processing equipment.

9. A computer readable storage medium storing computer instructions, wherein when the computer reads the computer instructions in the storage medium, the computer performs the intelligent recommendation method according to any one of claims 1 to 7.