CN117421355A - Search recall method, device and equipment - Google Patents
Search recall method, device and equipment Download PDFInfo
- Publication number
- CN117421355A CN117421355A CN202311469012.4A CN202311469012A CN117421355A CN 117421355 A CN117421355 A CN 117421355A CN 202311469012 A CN202311469012 A CN 202311469012A CN 117421355 A CN117421355 A CN 117421355A
- Authority
- CN
- China
- Prior art keywords
- information
- user
- search
- ordering
- recall
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 229940127554 medical product Drugs 0.000 claims abstract description 57
- 239000000825 pharmaceutical preparation Substances 0.000 claims abstract description 37
- 229940127557 pharmaceutical product Drugs 0.000 claims abstract description 37
- 230000006399 behavior Effects 0.000 claims description 22
- 238000004458 analytical method Methods 0.000 claims description 17
- 230000004044 response Effects 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000012163 sequencing technique Methods 0.000 abstract description 5
- 239000003814 drug Substances 0.000 description 7
- 230000011218 segmentation Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 4
- 239000008186 active pharmaceutical agent Substances 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 229940079593 drug Drugs 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012559 user support system Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/248—Presentation of query results
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2457—Query processing with adaptation to user needs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
- G06F18/2113—Selection of the most significant subset of features by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a search recall method, a device and equipment, wherein the method comprises the following steps: responding to the search request, and determining request information and user information; determining target data based on the request information and the user information, the target data including data describing user search preferences for the pharmaceutical product in different dimensions; obtaining a recall product set comprising first ordering information for each pharmaceutical product; calculating second ordering information for each medical product based on the target data and the first ordering information, wherein the first ordering information and the second ordering information are used for representing the demand and satisfaction of the user for the medical product, and the demand and satisfaction corresponding to the second ordering information are higher than those of the first ordering information; and updating the arrangement sequence of the medical products in the recall product set based on the second ordering information, and generating a search result. The search recall method provided by the invention can carry out secondary sequencing on search recall results of medical products, improves sequencing precision, and enables sequencing results to be closer to actual demands of users.
Description
Technical Field
The embodiment of the invention relates to the technical field of medicine product searching, in particular to a searching recall method, a searching recall device and searching recall equipment.
Background
Currently, in the field of medical search engines, when a user inputs information into the search engine, such as searching by using short words, the search engine responds to the input of the user to perform search recall, but if only TF-IDF (Term Frequency-inverse document Frequency) which is common to the search engine is applied in the search recall process, the TF is a weighting technique used for information retrieval and data mining. That is, the current search recall results do not reflect the user's actual preferences, and are less than optimal for the user's reference value.
Disclosure of Invention
The invention provides a search recall method, a device and equipment which can carry out secondary sequencing on search recall results of medical products, remarkably improve sequencing precision and enable the medical products provided preferentially to be closer to actual demands of users.
In order to solve the above technical problems, an embodiment of the present invention provides a search recall method, including:
responding to a search request of a user for a medical product, and determining request information and user information;
determining target data based on the request information and user information, the target data including data describing user search preferences for medical products in different dimensions;
obtaining a set of recall products generated in response to the search request, each pharmaceutical product in the set of recall products having first ordering information, the pharmaceutical products ordered based on the first ordering information;
calculating second ordering information for each medical product based on the target data and the first ordering information, wherein the first ordering information and the second ordering information are used for representing the demand and satisfaction of a user on the medical product, and the demand and satisfaction corresponding to the second ordering information are higher than those of the first ordering information;
and updating the medical product arrangement sequence in the recall product set based on the second ordering information, and generating a search result.
In some embodiments, the determining request information and user information includes:
determining login information of a user, browsing information after the login and input search information, wherein the search information is at least related to one or more of a pharmaceutical product name, a category, a function, a component, a brand, a tabu item, a price, a vending place, distribution information and a promotion;
parameter analysis is carried out on the browsing information and the searching information, and query conditions and ordering requirements are determined, wherein the query conditions are used for searching to obtain a recall product set;
and determining the identity and the geographic position of the user based on the login information.
In some embodiments, the method further comprises:
judging whether the sorting requirement meets a target triggering condition, if so, starting a personalized sorting mode to determine the target data, and responding to the sorting requirement by combining the target data;
if not, a default ordering mode is initiated in response to the ordering requirement.
In some embodiments, the method further comprises:
obtaining historical data of a target user, wherein the historical data comprises historical search information, historical browsing information and historical consumption data input by the user;
analyzing different dimensional characteristics in the historical data to obtain an analysis result, wherein the analysis result can represent offline multidimensional interest preference of a user on medical products;
and matching and storing the analysis result and the identity of the user to form a first data pool.
In some embodiments, the method further comprises:
recording near real-time browsing behaviors of a user and medical product information corresponding to the browsing behaviors, wherein the browsing behaviors comprise at least one of the following operations: purchasing, collecting, searching, sharing, clicking to view and comment;
calculating the near real-time multidimensional interest preference of the user based on the browsing behavior and commodity information;
and matching and storing the near-real-time multidimensional interest preference and the identity of the user to form a second data pool.
In some embodiments, the determining target data based on the request information and user information includes:
acquiring corresponding target offline multidimensional interest preference in a first data pool based on the identity;
obtaining corresponding target near-real-time multidimensional interest preference in a second data pool based on the identity;
and determining target data related to the request information in the target offline multidimensional interest preference degree and the target near-real-time multidimensional interest preference degree.
In some embodiments, the method further comprises:
determining relevance information of the pharmaceutical product and request information based on the first ordering information;
respectively configuring weight information for the relevance information and offline interest preference degree and near real-time interest preference degree of each dimension in target data based on priori data;
and constructing a fine ranking model for calculating the second ranking information based on the relevance information, the target data and the configured weight information.
In some embodiments, the calculating second ranking information for each of the pharmaceutical products based on the target data, first ranking information comprises:
and processing the target data and the first ordering information based on the fine ordering model to obtain second ordering information corresponding to each medical product.
Another embodiment of the present invention also provides a search recall apparatus, including:
the response module is used for responding to a search request of a user for the medical product and determining request information and user information;
a first determining module for determining target data according to the request information and the user information, the target data including data describing search preferences of the user for medical products in different dimensions;
an obtaining module for obtaining a set of recall products generated in response to the search request, each pharmaceutical product in the set of recall products having first ordering information, the pharmaceutical products being ordered based on the first ordering information;
the first calculation module is used for calculating second ordering information for each medical product according to the target data and the first ordering information, the first ordering information and the second ordering information are used for representing the demand and satisfaction of the user on the medical products, and the demand and satisfaction corresponding to the second ordering information are higher than those of the first ordering information;
and the sorting module is used for updating the arrangement sequence of the medical products in the recall product set according to the second sorting information and generating a search result.
Another embodiment of the present invention also provides an electronic device, including:
at least one processor; the method comprises the steps of,
a memory communicatively coupled to the at least one processor;
wherein the memory stores instructions executable by the at least one processor, the instructions configured to perform a search recall method as described in any one of the embodiments above.
Based on the disclosure of the above embodiment, it can be known that the embodiment of the present invention has the advantages that the request information and the user information are determined according to the search request input by the user, then the target data for describing the search preference of the user for the pharmaceutical product in different dimensions is determined based on the request information and the user information, the recall product set generated in response to the search request is obtained, then the second ranking information is calculated for each pharmaceutical product based on the target data and the first ranking information of each pharmaceutical product in the recall product set, the first ranking information and the second ranking information are both used for characterizing the demand and satisfaction degree of the user for the pharmaceutical product, the calculated demand and satisfaction degree of the second ranking information are higher than the first ranking information, and finally the pharmaceutical product ranking order in the recall product set is updated based on the second ranking information, so as to generate the search result. Through the process, the medical products in the recall product set undergo secondary fine ranking in combination with the interest preference of the user on the medical products, so that the sorting result is integrally and more fit with the actual demands of the user, the product finding experience of the user is improved, the shopping path of the user is shortened, and the shopping efficiency is improved.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. The objectives and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the present application is described in further detail below through the accompanying drawings and examples.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a search recall method in an embodiment of the invention.
FIG. 2 is a schematic application flow diagram of a search recall method in an application embodiment of the invention.
FIG. 3 is a schematic block diagram of a search recall apparatus in another embodiment of the invention.
Detailed Description
Hereinafter, specific embodiments of the present invention will be described in detail with reference to the accompanying drawings, but not limiting the invention.
It should be understood that various modifications may be made to the embodiments disclosed herein. Therefore, the following description should not be taken as limiting, but merely as exemplification of the embodiments. Other modifications within the scope and spirit of this disclosure will occur to persons of ordinary skill in the art.
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the application and, together with a general description of the disclosure given above and the detailed description of the embodiments given below, serve to explain the principles of the application.
These and other characteristics of the invention will become apparent from the following description of a preferred form of embodiment, given as a non-limiting example, with reference to the accompanying drawings.
It is also to be understood that, although the present application has been described with reference to some specific examples, a person skilled in the art will certainly be able to achieve many other equivalent forms of the invention, having the characteristics as set forth in the claims and hence all coming within the field of protection defined thereby.
The foregoing and other aspects, features, and advantages of the present application will become more apparent in light of the following detailed description when taken in conjunction with the accompanying drawings.
Specific embodiments of the present disclosure will be described hereinafter with reference to the accompanying drawings; however, it is to be understood that the disclosed embodiments are merely examples of the disclosure, which may be embodied in various forms. Well-known and/or repeated functions and constructions are not described in detail to avoid obscuring the disclosure in unnecessary or unnecessary detail. Therefore, specific structural and functional details disclosed herein are not intended to be limiting, but merely serve as a basis for the claims and as a representative basis for teaching one skilled in the art to variously employ the present disclosure in virtually any appropriately detailed structure.
The specification may use the word "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which may each refer to one or more of the same or different embodiments in accordance with the disclosure.
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
As shown in fig. 1, an embodiment of the present invention provides a search recall method, including:
s1: responding to a search request of a user for a medical product, and determining request information and user information;
s2: determining target data based on the request information and user information, the target data including data describing user search preferences for medical products in different dimensions;
s3: obtaining a set of recall products generated in response to the search request, each pharmaceutical product in the set of recall products having first ordering information, the pharmaceutical products ordered based on the first ordering information;
s4: calculating second ordering information for each medical product based on the target data and the first ordering information, wherein the first ordering information and the second ordering information are used for representing the demand and satisfaction of a user on the medical product, and the demand and satisfaction corresponding to the second ordering information are higher than those of the first ordering information;
s5: and updating the medical product arrangement sequence in the recall product set based on the second ordering information, and generating a search result.
Based on the foregoing, the solution of this embodiment is to determine request information and user information through a search request input by a user, then determine, based on the request information and the user information, target data for describing search preferences of the user for the pharmaceutical products in different dimensions, obtain a recall product set generated in response to the search request, and then calculate, based on the target data and first ranking information of each pharmaceutical product in the recall product set, second ranking information for each pharmaceutical product, where the first ranking information and the second ranking information are both used to characterize the demand and satisfaction of the user for the pharmaceutical product, and the calculated second ranking information corresponds to a demand and satisfaction higher than the first ranking information, and the final system updates the pharmaceutical product ranking order in the recall product set based on the second ranking information to generate a search result. Through the process, the medical products in the recall product set are subjected to secondary fine ranking by combining the interest preference of the user on the medical products, so that the sorting result is integrally and more fit with the actual demands of the user, namely, the method of the embodiment fuses the interest probability distribution sensing function of the user on the key attributes of the commodities on the basis of guaranteeing the commodity correlation, so that the secondary sorting realized on the basis of the function can enable the search result to be more intelligent and more fit with the shopping interest preference of the user, the effect required by the user is realized, the possibility of multiple screening of the user is reduced, the searching experience of the user is improved, the shopping path of the user is shortened, and the shopping efficiency is improved.
The search recall method in the embodiment can be applied to a search operation system of a medical platform or other similar or related type platforms, and can directly act on a search list meeting the sorting condition to realize the differentiated high-precision sorting of thousands of people and thousands of faces.
In an embodiment, the determining the request information and the user information includes:
s6: determining login information of a user, browsing information after the login and input search information, wherein the search information is at least related to one or more of a pharmaceutical product name, a category, a function, a component, a brand, a tabu item, a price, a vending place, distribution information and a promotion;
s7: parameter analysis is carried out on the browsing information and the searching information, and query conditions and ordering requirements are determined, wherein the query conditions are used for searching to obtain a recall product set;
s8: and determining the identity and the geographic position of the user based on the login information.
The method comprises the steps that login information of a user is determined to obtain identity information and login location information of the user, and browsing information after login is used for determining which medical products or related information in a commodity interface are browsed by the user and determining near-real-time multidimensional interest preference. The search information is determined to determine the medicine purchasing requirement of the current user, namely the medicine purchasing requirement of the current user. The search information may be related to information such as a pharmaceutical product name, category, brand, category, function, composition, brand, contra item, price, place of sale, distribution information, promotional program, etc., and is specifically indefinite. The search information can be input through a search input box in the interface and various related configuration options, meanwhile, the search information also comprises ordering requirement information, and a user can also input the ordering requirement information through the options configured by the interface. Based on the browsing information and the searching information, the system can perform parameter analysis, determine query conditions and ordering requirements, such as determining keywords of medicine names, brands and the like, determine query conditions, and determine ordering requirements based on an ordering mode selected by a user, such as ordering according to sales, ordering according to distance between delivery places, or default ordering. The system may search for a set of recall products based on the determined query conditions.
Because the scheme of the embodiment needs to rank the recall products in combination with the shopping interest preference of the user, whether the ranking requirement input by the user supports the system to reorder the recall products is conditional and can be implemented under the condition that the ranking requirement meets the requirement. Thus, the method of the present embodiment further comprises:
s9: judging whether the sorting requirement meets a target triggering condition, if so, starting a personalized sorting mode to determine the target data, and responding to the sorting requirement by combining the target data;
s10: if not, a default ordering mode is initiated in response to the ordering requirement.
For example, if the ordering requirement has no explicit indication, including distance, price, etc., the system may combine the shopping preferences of the user to secondarily order the products in the recalled product set.
In other embodiments, the method further comprises:
s11: obtaining historical data of a target user, wherein the historical data comprises historical search information, historical browsing information and historical consumption data input by the user;
s12: analyzing different dimensional characteristics in the historical data to obtain an analysis result, wherein the analysis result can represent offline multidimensional interest preference of a user on medical products;
s13: and matching and storing the analysis result and the identity of the user to form a first data pool.
Meanwhile, the method further comprises the following steps:
s14: recording near real-time browsing behaviors of a user and medical product information corresponding to the browsing behaviors, wherein the browsing behaviors comprise at least one of the following operations: purchasing, collecting, searching, sharing, clicking to view and comment;
s15: calculating the near real-time multidimensional interest preference of the user based on the browsing behavior and commodity information;
s16: and matching and storing the near-real-time multidimensional interest preference and the identity of the user to form a second data pool.
Specifically, the platform may record browsing and shopping behaviors, i.e., historical search information, historical browsing information, and historical consumption data, each time during the user history, where the browsing information includes browsed commodities, browsing duration of the same type of commodities, and other behaviors such as purchasing, collecting, searching, sharing, clicking for viewing, and commodity comments. And then, the stored data are scheduled periodically or periodically, and the interest degree data of the user in different dimensions, such as commodity brands, functions, prices and the like, are calculated through a calculation model with statistical analysis and information mining technologies, and are defined as offline multidimensional interest preference degrees in the embodiment and characterized as interest preference degrees of the user in the corresponding user historical browsing period calculated by the system after the user goes offline. After calculating interest preference degrees of a plurality of different dimensions, the system pushes the calculated data result set to a K-V memory, such as a Redis cluster, to form a user interest degree model data pool, namely a first data pool. Meanwhile, in order to determine the current interest preference of the user, the system monitors and records browsing behaviors of the user after entering the system program, including real-time operation behaviors of clicking (characterizing clicking to view commodities), purchasing, sharing, collecting and commenting on commodities, and the like of the user, wherein the behaviors of the user in single browsing are larger than a threshold value, then the system correspondingly determines commodities and commodity attributes corresponding to the browsing behaviors, and rapidly calculates by combining a pre-configured near-real-time user interest calculation formula with obtained data, and pushes a result characterizing the near-real-time multidimensional interest preference of the user to a K-V storage medium (such as a redis cluster) to form a second model pool.
The first model pool and the second model pool store a large number of interest preference data of different users, and when related data are required to be called, the models can provide matched interest preference data only by inputting the identity of the user into the models.
Further, the system determines target data based on the request information and the user information, and comprises:
s17: acquiring corresponding target offline multidimensional interest preference in a first data pool based on the identity;
s18: obtaining corresponding target near-real-time multidimensional interest preference in a second data pool based on the identity;
s19: and determining target data related to the request information in the target offline multidimensional interest preference degree and the target near-real-time multidimensional interest preference degree.
That is, the system may extract offline multidimensional interest preference and near real-time interest preference of corresponding users in the first data pool and the second data pool, and extract multidimensional interest preference data with higher correlation, i.e. higher reference value, from the interest preference data representing different periods of the users based on the request information of this time, including the above-mentioned search information and other self-extracted different dimensions, as target data, so as to further improve the sorting accuracy, and reduce the calculation load.
In other embodiments, to expedite the secondary ordering, the method further comprises:
s20: determining relevance information of the pharmaceutical product and request information based on the first ordering information;
s21: respectively configuring weight information for the relevance information and offline interest preference degree and near real-time interest preference degree of each dimension in target data based on priori data;
s22: and constructing a fine ranking model for calculating the second ranking information based on the relevance information, the target data and the configured weight information.
For example, the first ranking information in this embodiment is score (q, d) calculated by the TF-IDF scoring device for each drug in the recall product set, and the relevance information of the drug and the request information is expressed based on the score (q, d). The system then carries out weight configuration on the relevance information and the interest preference information of the user based on the obtained prior data, in this embodiment, the weight of the relevance information is set to a fixed value, the weight of the interest preference is changed, and the weight of the near-real-time interest preference is higher than that of the offline multidimensional interest preference, so that the system is more aware of the interests and the demands of the current user during sorting, and the sorting result is more suitable for the current interests and demands of the user. After determining the respective weights, the system may generate a refined model for computing the second ranking information based on the preconfigured calculation logic:
score (q, d) relevance weight + sigmoid Σ (offline interest preference degree offline weight) +Σ (near real-time interest preference degree near real-time weight)
Based on the formula, the system can calculate and obtain the personalized comprehensive score of each medical product in the recall product set, wherein the score is the second ordering information, and the higher the score is, the higher the fit degree between the actual interest and the demand of the characterization and the user is.
After the fine ranking model is generated, when the system calculates the second ranking information for each medical product based on the target data and the first ranking information, the system can process the target data and the first ranking information based on the fine ranking model to obtain the second ranking information corresponding to each medical product.
To better illustrate the above embodiments, the following description of the overall application flow is provided in conjunction with the specific application and fig. 2:
as shown in fig. 2, the method in the present embodiment is applied to a medical platform equipped with a personalized ranking component for performing the method of the embodiments of the present application. In particular, the method comprises the steps of,
1. the user enters application service, keywords are input into a search interface to search, and the system enters a comprehensive search ordering state by default;
2. after receiving a search request input by a user, the search API analyzes parameters to obtain key parameters, wherein the key parameters comprise longitude and latitude corresponding to a login ID of the user, keywords input by the user, identity identification of the user and the like. The longitude and latitude can be specifically the longitude and latitude of the GPS positioning of the mobile terminal of the user, and are used for defining the geographic position of the user and providing medicine providing and distributing services for the user by adapting the merchant which can provide services within the LBS geographic range of the user;
3. through the word segmentation device of the search query end, the word segmentation processing can be specifically performed on search requests input by users based on the custom word stock with word parts and NLP word segmentation device (such as Ansj, bargained word segmentation, IK and the like) in the pharmaceutical industry, and the entity recognition processing is performed based on word segmentation results, namely, corresponding pharmaceutical products are roughly determined;
4. performing service retrieval condition assembly and splicing on the parameter objects subjected to comprehensive processing of the request parameters to form a query data structure conforming to search engine rules; for example, keywords input by a user, GPS location information of the user, ordering requirements, other advanced screening indexes (such as three-level classification identifier of a pharmaceutical product, brand), single page data display amount, implicit service operation identifier and the like are comprehensively processed to form parameter objects containing various result information for searching references by a search engine, ordering programs and the like to be referenced, such as identity identifiers of the user, word segmentation results, predicted entity objects, ordering requirements, merchants capable of providing services and the like.
5. The system sends a request to a search engine cluster through a client of the search engine based on the query data structure;
6. the search engine cluster (such as an open source search engine (SolrCloud, elasticsearch)) receives the request and then analyzes the request data structure to determine the query conditions, ordering requirements and identity.
7. The comprehensive ordering custom ordering component obtains the query condition, ordering requirement and identity mark, and determines ordering strategy based on ordering requirement:
strategy 1: when the request parameters have the identity marks conforming to the rules and the specified ordering requirement is comprehensive ordering, the dynamic multi-thread routing is started to execute the following tasks:
thread 1: recall the commodity meeting the condition rule by using the query condition, and calculate score (q, d) for each medical product in the recalled commodity set according to the designated TF-IDF scoring device; thread 2: acquiring offline user multidimensional interest preference data of a user from an offline user interest model pool by using an identity;
thread 3: acquiring multidimensional interest preference data of the user near-real-time user from a near-real-time user interest model pool by using the identity;
strategy 2: when the request parameters are comprehensive sorting and the user identification does not accord with the rule, only a single thread uses the query condition to recall the commodity meeting the rule, calculates score (q, d) of the corresponding commodity for the recalled commodity set according to the designated TF-IDF scoring device, and sorts the commodity based on the score (q, d);
8. after receiving the recall commodity set, the offline user multidimensional interest preference and the near real-time user interest preference, the personalized sorting device in the comprehensive sorting custom sorting component uses a custom personalized sorting model to perform data matching and score model fusion on three data sets, finally obtains personalized comprehensive scores of each commodity, and performs inverted ranking according to the scores from high to low:
the multidimensional interest preference mainly comprises preference of users for products, preference of price intervals, promotion preference, brand preference and the like. The multi-dimensional interest preference of the offline user is consistent with the structure of the near-real-time interest preference data, a personalized secondary fine-ranking fusion model is built by combining the weight ratio of the pre-configured offline, near-real-time and correlation, and the personalized comprehensive score is calculated by using the model:
score (q, d) relevance weight + sigmoid Σ (offline interest preference degree offline weight) +Σ (near real-time interest preference degree near real-time weight)
9. Returning to a search API according to the recall result of the comprehensive personalized secondary fine arrangement;
10. and the searching API performs service processing on the result after secondary fine discharge according to the service rule, encapsulates the commodity attribute and returns to the final request terminal.
As shown in fig. 3, another embodiment of the present invention also provides a search recall apparatus 100, including:
the response module is used for responding to a search request of a user for the medical product and determining request information and user information;
a first determining module for determining target data according to the request information and the user information, the target data including data describing search preferences of the user for medical products in different dimensions, the medical products being ordered based on the first ordering information;
an obtaining module for obtaining a set of recall products generated in response to the search request, each pharmaceutical product in the set of recall products having first ordering information;
the first calculation module is used for calculating second ordering information for each medical product according to the target data and the first ordering information, the first ordering information and the second ordering information are used for representing the demand and satisfaction of the user on the medical products, and the demand and satisfaction corresponding to the second ordering information are higher than those of the first ordering information;
and the sorting module is used for updating the arrangement sequence of the medical products in the recall product set according to the second sorting information and generating a search result.
In some embodiments, the determining request information and user information includes:
determining login information of a user, browsing information after the login and input search information, wherein the search information is at least related to one or more of a pharmaceutical product name, a category, a function, a component, a brand, a tabu item, a price, a vending place, distribution information and a promotion;
parameter analysis is carried out on the browsing information and the searching information, and query conditions and ordering requirements are determined, wherein the query conditions are used for searching to obtain a recall product set;
and determining the identity and the geographic position of the user based on the login information.
In some embodiments, the method further comprises:
judging whether the sorting requirement meets a target triggering condition, if so, starting a personalized sorting mode to determine the target data, and responding to the sorting requirement by combining the target data;
if not, a default ordering mode is initiated in response to the ordering requirement.
In some embodiments, the method further comprises:
obtaining historical data of a target user, wherein the historical data comprises historical search information, historical browsing information and historical consumption data input by the user;
analyzing different dimensional characteristics in the historical data to obtain an analysis result, wherein the analysis result can represent offline multidimensional interest preference of a user on medical products;
and matching and storing the analysis result and the identity of the user to form a first data pool.
In some embodiments, the method further comprises:
recording near real-time browsing behaviors of a user and medical product information corresponding to the browsing behaviors, wherein the browsing behaviors comprise at least one of the following operations: purchasing, collecting, searching, sharing, clicking to view and comment;
calculating the near real-time multidimensional interest preference of the user based on the browsing behavior and commodity information;
and matching and storing the near-real-time multidimensional interest preference and the identity of the user to form a second data pool.
In some embodiments, the determining target data based on the request information and user information includes:
acquiring corresponding target offline multidimensional interest preference in a first data pool based on the identity;
obtaining corresponding target near-real-time multidimensional interest preference in a second data pool based on the identity;
and determining target data related to the request information in the target offline multidimensional interest preference degree and the target near-real-time multidimensional interest preference degree.
In some embodiments, the method further comprises:
determining relevance information of the pharmaceutical product and request information based on the first ordering information;
respectively configuring weight information for the relevance information and offline interest preference degree and near real-time interest preference degree of each dimension in target data based on priori data;
and constructing a fine ranking model for calculating the second ranking information based on the relevance information, the target data and the configured weight information.
In some embodiments, the calculating second ranking information for each of the pharmaceutical products based on the target data, first ranking information comprises:
and processing the target data and the first ordering information based on the fine ordering model to obtain second ordering information corresponding to each medical product.
Another embodiment of the present invention also provides an electronic device, including:
at least one processor; the method comprises the steps of,
a memory communicatively coupled to the at least one processor;
wherein the memory stores instructions executable by the at least one processor, the instructions configured to perform a search recall method as described in any one of the embodiments above.
Another embodiment of the present invention also provides a storage medium including a stored program, wherein the program, when run, controls a device including the storage medium to perform the search recall method according to any one of the embodiments described above.
Embodiments of the present invention also provide a computer program product tangibly stored on a computer-readable medium and comprising computer-readable instructions that, when executed, cause at least one processor to perform a search recall method such as in the embodiments described above. It should be understood that each solution in this embodiment has a corresponding technical effect in the foregoing method embodiment, which is not described herein.
It should be noted that, the computer storage medium of the present application may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage media element, a magnetic storage media element, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present application, however, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, antenna, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
The above embodiments are only exemplary embodiments of the present invention and are not intended to limit the present invention, the scope of which is defined by the claims. Various modifications and equivalent arrangements of this invention will occur to those skilled in the art, and are intended to be within the spirit and scope of the invention.
Claims (10)
1. A search recall method, comprising:
responding to a search request of a user for a medical product, and determining request information and user information;
determining target data based on the request information and user information, the target data including data describing user search preferences for medical products in different dimensions;
obtaining a set of recall products generated in response to the search request, each pharmaceutical product in the set of recall products having first ordering information, the pharmaceutical products ordered based on the first ordering information;
calculating second ordering information for each medical product based on the target data and the first ordering information, wherein the first ordering information and the second ordering information are used for representing the demand and satisfaction of a user on the medical product, and the demand and satisfaction corresponding to the second ordering information are higher than those of the first ordering information;
and updating the medical product arrangement sequence in the recall product set based on the second ordering information, and generating a search result.
2. The search recall method of claim 1, wherein determining request information and user information comprises:
determining login information of a user, browsing information after the login and input search information, wherein the search information is at least related to one or more of a pharmaceutical product name, a category, a function, a component, a brand, a tabu item, a price, a vending place, distribution information and a promotion;
parameter analysis is carried out on the browsing information and the searching information, and query conditions and ordering requirements are determined, wherein the query conditions are used for searching to obtain a recall product set;
and determining the identity and the geographic position of the user based on the login information.
3. The search recall method of claim 2, wherein the method further comprises:
judging whether the sorting requirement meets a target triggering condition, if so, starting a personalized sorting mode to determine the target data, and responding to the sorting requirement by combining the target data;
if not, a default ordering mode is initiated in response to the ordering requirement.
4. The search recall method of claim 2, wherein the method further comprises:
obtaining historical data of a target user, wherein the historical data comprises historical search information, historical browsing information and historical consumption data input by the user;
analyzing different dimensional characteristics in the historical data to obtain an analysis result, wherein the analysis result can represent offline multidimensional interest preference of a user on medical products;
and matching and storing the analysis result and the identity of the user to form a first data pool.
5. The search recall method of claim 4, further comprising:
recording near real-time browsing behaviors of a user and medical product information corresponding to the browsing behaviors, wherein the browsing behaviors comprise at least one of the following operations: purchasing, collecting, searching, sharing, clicking to view and comment;
calculating the near real-time multidimensional interest preference of the user based on the browsing behavior and commodity information;
and matching and storing the near-real-time multidimensional interest preference and the identity of the user to form a second data pool.
6. The search recall method of claim 5, wherein the determining target data based on the request information and user information comprises:
acquiring corresponding target offline multidimensional interest preference in a first data pool based on the identity;
obtaining corresponding target near-real-time multidimensional interest preference in a second data pool based on the identity;
and determining target data related to the request information in the target offline multidimensional interest preference degree and the target near-real-time multidimensional interest preference degree.
7. The search recall method of claim 6, further comprising:
determining relevance information of the pharmaceutical product and request information based on the first ordering information;
respectively configuring weight information for the relevance information and offline interest preference degree and near real-time interest preference degree of each dimension in target data based on priori data;
and constructing a fine ranking model for calculating the second ranking information based on the relevance information, the target data and the configured weight information.
8. The search recall method of claim 7 wherein the calculating second ranking information for each of the pharmaceutical products based on the target data, first ranking information comprises:
and processing the target data and the first ordering information based on the fine ordering model to obtain second ordering information corresponding to each medical product.
9. A search recall device, comprising:
the response module is used for responding to a search request of a user for the medical product and determining request information and user information;
a first determining module for determining target data according to the request information and the user information, the target data including data describing search preferences of the user for medical products in different dimensions;
an obtaining module for obtaining a set of recall products generated in response to the search request, each pharmaceutical product in the set of recall products having first ordering information, the pharmaceutical products being ordered based on the first ordering information;
the first calculation module is used for calculating second ordering information for each medical product according to the target data and the first ordering information, the first ordering information and the second ordering information are used for representing the demand and satisfaction of the user on the medical products, and the demand and satisfaction corresponding to the second ordering information are higher than those of the first ordering information;
and the sorting module is used for updating the arrangement sequence of the medical products in the recall product set according to the second sorting information and generating a search result.
10. An electronic device, comprising:
at least one processor; the method comprises the steps of,
a memory communicatively coupled to the at least one processor;
wherein the memory stores instructions executable by the at least one processor, the instructions configured to perform the search recall method of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311469012.4A CN117421355A (en) | 2023-11-07 | 2023-11-07 | Search recall method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311469012.4A CN117421355A (en) | 2023-11-07 | 2023-11-07 | Search recall method, device and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117421355A true CN117421355A (en) | 2024-01-19 |
Family
ID=89522693
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311469012.4A Pending CN117421355A (en) | 2023-11-07 | 2023-11-07 | Search recall method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117421355A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117786242A (en) * | 2024-02-26 | 2024-03-29 | 腾讯科技（深圳）有限公司 | Searching method based on position and related device |
-
2023
- 2023-11-07 CN CN202311469012.4A patent/CN117421355A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117786242A (en) * | 2024-02-26 | 2024-03-29 | 腾讯科技（深圳）有限公司 | Searching method based on position and related device |
| CN117786242B (en) * | 2024-02-26 | 2024-05-28 | 腾讯科技（深圳）有限公司 | Searching method based on position and related device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11449919B2 (en) | Commodity recommendation method and commodity recommendation device | |
| US11995112B2 (en) | System and method for information recommendation | |
| CN109033101B (en) | Label recommendation method and device | |
| US20160191450A1 (en) | Recommendations Engine in a Layered Social Media Webpage | |
| US20140379617A1 (en) | Method and system for recommending information | |
| US20150242750A1 (en) | Asymmetric Rankers for Vector-Based Recommendation | |
| US20150186938A1 (en) | Search service advertisement selection | |
| CN110135951B (en) | Game commodity recommendation method and device and readable storage medium | |
| CN107832338B (en) | Method and system for recognizing core product words | |
| CN110111167A (en) | A kind of method and apparatus of determining recommended | |
| CN111026853B (en) | Target problem determining method and device, server and customer service robot | |
| CN103309869A (en) | Method and system for recommending display keyword of data object | |
| CN113077317A (en) | Item recommendation method, device and equipment based on user data and storage medium | |
| CN117421355A (en) | Search recall method, device and equipment | |
| CN111310046A (en) | Object recommendation method and device | |
| CN116308684B (en) | Online shopping platform store information pushing method and system | |
| CN111078997B (en) | Information recommendation method and device | |
| CN114820123A (en) | Group purchase commodity recommendation method, device, equipment and storage medium | |
| CN114581175A (en) | Commodity pushing method and device, storage medium and electronic equipment | |
| US10296956B2 (en) | Method, system, and computer-readable medium for product and vendor selection | |
| CN115423555A (en) | Commodity recommendation method and device, electronic equipment and storage medium | |
| CN114547385A (en) | Label construction method and device, electronic equipment and storage medium | |
| US20230099627A1 (en) | Machine learning model for predicting an action | |
| US20210201262A1 (en) | Freight load matching system | |
| CN110827101A (en) | Shop recommendation method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |