CN110968801A - Real estate product searching method, storage medium and electronic device - Google Patents
Real estate product searching method, storage medium and electronic device Download PDFInfo
- Publication number
- CN110968801A CN110968801A CN201911226750.XA CN201911226750A CN110968801A CN 110968801 A CN110968801 A CN 110968801A CN 201911226750 A CN201911226750 A CN 201911226750A CN 110968801 A CN110968801 A CN 110968801A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- product
- matching
- property product
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9538—Presentation of query results
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for searching a local product, a storage medium and electronic equipment, wherein the local product is preset with a state attribute with multilevel weight, and the method comprises the following steps: obtaining a search request, wherein the search request comprises a user intention of a multi-level weight; matching the local product with the search request according to the user intention and the state attribute to obtain a matching result; sorting the matching results by using a sorting model to obtain sorting results, wherein the sorting model is obtained by adopting multi-level weight training data training; and obtaining the target property product according to the sequencing result. Because the user intention and the real estate product both have multi-level weight, matching can be performed according to the intention of different weights in the user intention and the state attributes of different weights in the state attributes of the real estate product, multi-level matching of the real estate product and a search request can be achieved, matching is performed in all aspects, people and houses are matched more accurately, recalled house source information meets the current requirements of the user, and the user requirements are located quickly and accurately.
Description
Technical Field
The invention relates to the technical field of data search, in particular to a method for searching a local product, a storage medium and electronic equipment.
Background
With the rapid development of the internet technology, people's lives are more and more convenient, and Location-Based services (LBS) enable users to conveniently acquire various information around the positions, so that user experience is greatly improved. The LBS is a value added service that obtains the location information (geographic coordinates) of the Mobile terminal user through the radio communication network of the Mobile operator (such as Global System for Mobile Communications, GSM), Long Term Evolution (Long Term Evolution, LTE) network) or external Positioning (such as Global Positioning System (GPS)), and provides corresponding services to the user with the support of the GIS platform. In the real estate field, such as renting or selling houses, for example, when a user has a house renting requirement, the user can inquire the house sources of the stored target area according to the area where the user is located, then all the house sources meeting the conditions are displayed to the user, and the user selects the house sources expected by the user according to the requirement.
However, in current real estate products, such as long rental apartment fields, because of too many house renting factors including traffic conditions, social place information, commute time, distance, cell facilities, house configuration, and the like, because the current LBS search method can only obtain location-related resources and information, for more complex real estate products, the search results obtained based on the LBS search method are difficult to meet the needs of users.
Disclosure of Invention
The technical scheme provides a method for searching a local product, a storage medium and electronic equipment, so as to obtain a search result which is more in line with the requirements of users.
According to a first aspect, an embodiment of the present invention provides a method for searching a local product, where the local product has preset state attributes with multiple levels of weights, and the method includes: obtaining a search request, wherein the search request comprises a user intention of a multi-level weight; matching the local product with the search request according to the user intention and the state attribute to obtain a matching result; sorting the matching results by using a sorting model to obtain sorting results, wherein the sorting model is obtained by adopting multi-level weight training data training; and obtaining the target property product according to the sequencing result.
Optionally, the search request includes multidimensional data information; matching the local product with the search request according to the user intention and the state attribute to obtain a matching result, wherein the matching result comprises the following steps: respectively recalling the real estate products according to preset recalling rules corresponding to different dimensional data information based on the multi-dimensional data information, and establishing a plurality of recall index sets which respectively correspond to the multi-dimensional data one by one; performing matching calculation on the real estate products and the search requests in the plurality of recall index sets hierarchically according to the user intention and the weight level of the state attribute to obtain the matching degree; and carrying out preliminary sorting on the real estate products according to the matching degree to obtain a preliminary sorting result.
Optionally, the multidimensional data information includes: at least two of user demand information, user behavior information, user portrait information, and real estate product portrait information; the recall index set includes: at least two of a search index collection, a behavior preference index collection, a user portrait index collection, a property product heat index collection.
Optionally, the method for establishing the search index set includes: recalling the user demand information by using a log analysis frame, and establishing a search index set; the user behavior information includes: the method for establishing the behavior preference index set comprises the following steps: recalling the behavior of the user based on the historical behavior information and the real-time behavior information of the user, and establishing a behavior preference index set; the method for establishing the user portrait index set comprises the following steps: acquiring a user portrait; mining the real estate product information corresponding to the user image in a preset real estate product database; screening real estate product information based on location service, and establishing a user portrait index set; the method for establishing the hot index set of the production product comprises the following steps: acquiring a property product database corresponding to the search request based on the search request; and screening the real estate product information with the popularity being larger than the preset value from a database corresponding to the search request, and establishing a real estate product popularity index set.
Optionally, the ranking the matching result by using a ranking model, and obtaining a ranking result includes: acquiring training data, wherein the training data comprises multi-level sample data and sample weight information, and the sample weight information comprises weight information of user intention and weight information of state attributes; training a preset sequencing model by using training data to obtain a trained sequencing model; and sequencing the recall results by utilizing the trained sequencing model to obtain a sequencing result.
Optionally, the multi-level sample data comprises multi-level online sample data and multi-level offline sample data.
Optionally, after obtaining the target property product according to the sorting result, the method includes: acquiring intention information of a supplier of a real estate product; the property product corresponding to the intention information of the property product supplier is added to the target property product according to the intention information of the property product supplier.
Optionally, obtaining the search request includes: acquiring a search character string input by a user; and performing at least one of recommendation, completion and consciousness recognition on the search character string to obtain a search request.
According to a second aspect, embodiments of the present invention provide a computer-readable storage medium having stored thereon computer instructions for causing a computer to execute the method of searching for a property product described in any one of the above first aspects.
According to a third aspect, an embodiment of the present invention provides an electronic device, including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to cause the at least one processor to perform the property product searching method as described in any one of the above first aspects.
Because the user intention and the real estate product both have multi-level weights, matching can be performed according to intentions with different weights in the user intention and state attributes with different weights in the state attributes of the real estate product, multi-level matching of the real estate product and a search request can be achieved, the main demand intention of the user is matched with the real estate product to which the main key attribute corresponding to the main demand intention belongs, the secondary demand intention is matched with the real estate product to which the secondary key attribute corresponding to the secondary demand intention belongs, the pertinence is stronger, matching strategies are performed on all aspects to enable people-house matching to be more accurate, recalled house source information meets the current requirements of the user, and user requirements are located quickly and accurately.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic diagram showing a property product searching method of the present embodiment;
FIG. 2 is a schematic diagram showing the method of matching a property product with a search request of the present embodiment;
fig. 3 shows a schematic view of an electronic device of an embodiment of the invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Currently, a property product can only obtain resources and information related to a position through LBS search, for example, in the field of long rented apartments, after the geographic position of an apartment is determined, the LBS search mode can provide various information services related to the geographic position of the apartment: for example, the geographic location of an apartment is at the XX position, and the system searches various related service information, such as traffic, leisure and entertainment places, shopping squares and other service circle information, within a preset area range based on the current location according to the current geographic location. However, currently, the factors considered by the user when searching for the real estate products are far more than the resources and information related to the location, for example, in the field of long-rented apartments, the long-rented house factors considered by the user include many aspects such as traffic conditions, price information, renting modes, social place information, commuting time, distance, cell facilities, house state structures, and the like, wherein the price information, the renting modes, and the house state structures are key attributes of the long-rented apartments. Therefore, the LBS search is difficult to cover key attributes of the long-rental apartment, and the current recall mode of the search in the field of the long-rental apartment is often based on the inverted index in combination with the recall of the user behavior, and the recall results are sorted only based on the recall results of the user behavior, which are all attributes that the long-rental apartment has and the weights of the attributes, such as price information, lease mode, house state structure, and the like, are not considered. This can result in recalls and ranked results that are difficult to match well with the user's search request. Therefore, the inventor weights the status attributes of the property products according to the influence factors of the property field, for example, the status attributes can be divided into a key status attribute and a non-key status attribute, and as an exemplary embodiment, the key status attribute can include at least one of location information, price information and rental mode of the property product. The non-critical status attributes may include at least one of room area information, orientation information, architectural style information, environmental information, and companion information. Based on this, an embodiment of the present invention provides a method for searching a property product, which may include the following steps, as shown in fig. 1:
s11, obtaining a search request. In this embodiment, the search request may be a search string input by the user for representing the text intended by the user, and the search request may be the text input by the user for describing the origin of the heart instrument, taking long rental apartment field as an example. For example, the search request may be XX building of XX city XX cell. Or XX district/region graduate preferential activities, etc. The search request may be any condition that characterizes the intention of the user and is arbitrarily input by the user, and is not limited to the exemplary example described in the above embodiment. Specifically, as an alternative embodiment, for example, the following may be performed for the search string: query recommendation, query completion and related search keywords. And identifying the intention of the user, and accurately positioning the intention of the user. For example, a user searches by directly filling in or collecting user requirement feedback by a system, for example, inputting two characters "facing sun" in a search box, and recommending search words such as "facing sun door" and "facing sun first house" in a real-time reinforcement learning manner; when the input is the 'Chaoyangyi' typeface, the system automatically queries and completes, and then identifies the intention of the user. In this embodiment, in the recognition of the user's intention, a Point of Interest (POI) dictionary category may be complemented, so that the user's intention may be accurately recognized, for example, in the field of long rental apartments, a digital city and a building dictionary library may be self-built and purchased. After completing the character string input by the user and recognizing the user intention, the user intention may be divided into a plurality of weighted user intentions according to conventional influence factors in the field of long rental apartments, for example, into a primary demand intention and a secondary demand intention, and the primary demand intention may include at least one of section information, price information, and rental style of the property product as an exemplary embodiment. The secondary demand intention may include at least one of room area information orientation information, architectural style information, environmental information, and supporting information. For example, the architectural style information may include information about whether there is a balcony, whether there is a transparency, and the shape of a room.
And S12, matching the local product with the search request according to the user intention and the state attribute to obtain a matching result. As an exemplary embodiment, after the user intention is identified, matching can be respectively carried out according to intentions with different weights in the user intention and state attributes with different weights in the state attributes of the property product, multi-level matching of search results can be realized, matching is carried out in a grading mode according to the criticality of the user intention and the property product attribute, the pertinence is stronger, and the matching degree of the user intention and the property product can be higher by utilizing the weights of the user intention and the state attribute weights of the property product to carry out corresponding matching. For example, in the field of long-rented apartments, because the long-rented apartments involve many factors, the long-rented apartments need to search for matching house resources, such as transportation, shopping, commuting distance, time, basic information in house resources, service information, and the like, according to the requirements of users in various aspects. In this embodiment, different weights may be preset for the influencing factors of the house source, for example, attributes such as position, price, rental mode, and the like may be used as primary key attributes, and attributes such as area, orientation, environment, and support are secondary key attributes, so that sectional multi-level matching may be performed according to the state attributes of different levels of weights, the primary demand intention of the user is matched with the property product to which the primary key attribute corresponding to the primary demand intention belongs, and the secondary demand intention is matched with the property product to which the secondary key attribute corresponding to the secondary demand intention belongs. The matching strategy is made in all aspects, so that the matching of the house is more accurate, the recalled house source information meets the current requirements of the user, and the user requirements are quickly and accurately positioned.
And S13, sequencing the matching results by utilizing a sequencing model to obtain a sequencing result, wherein the sequencing model is obtained by adopting multi-level weight training data training. As an exemplary embodiment, the ranking model may employ a ranking Learning (L2R) model to rank the matching results. In training the model, multi-level weight training data is used. In this embodiment, the multi-level weight training data may include multi-level sample data and sample weight information, where the sample weight information includes weight information of user intent and weight information of state attributes. Specifically, taking the field of long rental apartments as an example, the multi-level sample data may include online multi-level sample data and offline multi-level sample data, and for example, the online multi-level sample data may include: clicking, ordering and other user online operation information; the multi-level offline sample data can comprise offline field watching, reserving and signing, and requirements of users collected after watching, reserving and signing. In this embodiment, the sample weight information may be weights of different levels of samples in the multi-level sample data, such as weight information of the collected user behavior user representation and the house source representation.
Specifically, the results of the online operation and the offline operation can be used to label the multi-level online sample data and the multi-level offline sample data, and the labeled sample data is used to train the model.
Illustratively, the multi-level online sample data can be obtained by collecting user behaviors, user portraits and house source portraits, specifically, the user behaviors, the user portraits and the house source portraits can be collected by an ELK module of a data warehouse Hive, and a so-called ELK framework can include a search server Elasticsearch, a server-side data processing pipeline logstack and an open source analysis and visualization platform Kibana, wherein the logstack is used for collecting, analyzing and filtering logs, and the logs are stored in the Elasticsearch after being received from various sources. The Elasticissearch mainly stores data information such as historical behavior logs of users, user images, house source images and the like; kibana is a Web-based graphical interface for searching, analyzing and visualizing the log data stored in the Elasticissearch index and retrieving the data using the REST interface of the Elasticissearch. After user behaviors and portraits and house source portraits are collected, massive data information can be processed in batch by utilizing a Spark ecology, namely a Berkely data analysis stack, a data cluster which accords with the current user is obtained by carrying out characteristic engineering processing such as gathering, refining, screening, complementing and the like on multi-level online sample data and multi-level offline sample data, training data of an L2R model are mined and extracted from the data cluster, and the L2R model is trained by utilizing the training data.
Specifically, the real-time operation behaviors of each current user can be called when the sequencing model is used for sequencing, the real-time operation behaviors provided by the user side are uniformly collected by adopting a parallel loading mechanism, and then the real-time operation behaviors are cleaned in real time, and user real-time behavior information is analyzed and stored; and summarizing the real-time behavior information of the user and the data clusters which are obtained by utilizing Spark ecological processing and accord with the current user, inputting a sequencing model to perform matching sequencing scoring on the requirements of the user, and obtaining a sequencing result.
In this embodiment, the sorting result may be returned to the elv module of Hive, and the logstack analyzes and filters the returned log data, and stores the log data in the Elasticsearch to provide data support for the next training of the new model.
The following is an exemplary description of the training process of the ranking model:
assuming that given a query, the result list returned by the search is L, the set of documents clicked by the user is C, if one document di is clicked, the other document dj is not clicked, and dj is arranged before di in the result list, then di > dj is a training record; namely, the training data is: { di > dj | di belongs to C, dj belongs to L-C, p (dj) < p (di) }, wherein p (d) represents the position of the document d in the query result list, and smaller represents more forward, so as to finally obtain a real sequence; the algorithm in L2R is then used to learn a ranking model that outputs document sequences that are as similar as possible to the true sequences.
And S14, obtaining a target property product according to the sorting result.
As an exemplary embodiment, the search request includes multidimensional data information, the search request may be recalled in multiple channels for the multidimensional data, and recall results may be sorted in different dimensions, specifically, the step S12 matches the local product with the search request according to the user intention and the status attribute, and obtaining the matching result may include the following steps:
s121, the real estate products are recalled respectively according to preset recall rules corresponding to different dimensional data information based on the multi-dimensional data information, and a plurality of recall index sets are obtained. The plurality of recall index sets are respectively in one-to-one correspondence with the multidimensional data. In this embodiment, the multi-dimensional data may include at least one of user demand information, user behavior information, user portrait information, and target portrait information to be searched. The recall index set includes: at least one of a search index collection, a behavior preference index collection, a user representation index collection, a property product heat index collection. Illustratively, recalling data based on different dimensions respectively may include:
the method comprises the steps of for recalling user demand information, collecting user demand feedback through user filling or inquiry, obtaining the user demand information, processing the user demand information through a log analysis frame such as an ELK frame, exemplarily, obtaining the user demand information as 'zhang yang first house', selecting data information of a long-rented apartment near the 'zhang first house' from a house source library, transmitting the data to Logstash for filtering and formatting, then storing the data in an Elasticisearch, and establishing a search index set.
The recall of the user behavior information can recall the behavior of the user based on the historical behavior information and the real-time behavior information of the user, and a behavior preference index set is established. Specifically, data cleaning is carried out on the real-time behavior information; the real-time behavior information after data cleaning is processed in real time, in this embodiment, an open source stream processing framework, such as a Flink processing framework, may be used for real-time processing to obtain a real-time behavior processing result; since the basic data model of Flink is a sequence of data streams and events (events); in the stream execution mode, the output of an event after being processed in one node can be sent to the next node for immediate processing, and the execution engine does not introduce extra delay and can process the real-time line information of the user with lower delay. Meanwhile, historical behavior information is collected; processing the summarized historical behavior information in batches to obtain a historical behavior calculation result; in this embodiment, a Spark processing framework may be used to perform batch processing on the history information, and as Spark is used to generate the history behavior log, the generated history behavior log may be easily cut into smaller blocks to perform calculation processing, so as to achieve the effect of low delay. Summarizing a real-time behavior processing result and a historical behavior calculation result, and establishing a behavior preference index set;
recalling the user portrait information can mine the real estate product information corresponding to the user portrait in a preset database after the user portrait is obtained; and screening the real estate product information based on the LBS, and establishing a user portrait index set. Specifically, the user portrait information can be embodied, the property product information corresponding to the user portrait is mined according to the user portrait information, in the case of a long rental apartment, the characteristics of the user such as house source preference, surrounding preference and traffic condition can be judged according to the user portrait, property products relevant to the user preference are mined from a house source library, and a user portrait index set is established. Because the search request is recalled by adopting the user portrait information, the marketing can be accurately carried out on the user, and products preferred by the user are recommended.
For recalling the portrait information of the real estate product, a database of the real estate product corresponding to the search request can be obtained based on the search request, the information of the real estate product with the popularity larger than a preset value is screened out from the database corresponding to the search request, and a hot index set of the real estate product is established. For example, in the field of long-rent apartments, after the identification of the user intention is identified, a house source library related to the user intention can be searched according to the user intention, and high-heat long-rent apartment data can be mined from the obtained house source library, so that a real-estate product heat index set is established.
And S122, carrying out matching calculation on the property products and the search requests in the plurality of recall index sets in a hierarchical manner according to the user intention and the weight level of the state attribute to obtain the matching degree. For example, in the case of a long rental apartment, according to a "target apartment" searched by a user, LBS matching is first performed on a main demand intention of the user, for example, district information, and price matching can be performed on the price of the demand of the user and the price of the house source. And then, calculating the matching degree of the non-key state attributes of the house sources according to the secondary demand intentions of the user, such as room area information, orientation information, architectural style information, environment information, matching information and the like, and acquiring the matching degree information of each real estate product. And performing segmented multilevel matching according to the state attributes of different level weights, so that the main demand intention of the user can be matched with the property product to which the main key attribute corresponding to the main demand intention belongs, and the secondary demand intention can be matched with the property product to which the secondary key attribute corresponding to the secondary demand intention belongs. The matching strategy is made in all aspects, so that the matching of the house is more accurate, the recalled house source information more meets the current requirements of the user, and the user requirements are quickly and accurately positioned.
And S123, carrying out preliminary sorting on the real estate products according to the matching degree to obtain a preliminary sorting result.
And S124, taking the preliminary sorting result as a matching result.
By improving the personalized recall channel of the user and recalling the search request of the user in different dimensions, the recall channel is diversified, the number of recalled house resources is obviously increased, the range of the house resources is wider, hotspots and boutique house resources can be recalled, and the quality of the house resources is improved. And carrying out sectional multi-level matching according to the state attributes of different level weights, so that the main demand intention of the user can be matched with the property product to which the main key attribute corresponding to the main demand intention belongs, and the secondary demand intention can be matched with the property product to which the secondary key attribute corresponding to the secondary demand intention belongs. The matching strategy is made in all aspects, so that the matching of the house is more accurate, the recalled house source information meets the current requirements of the user, and the user requirements are quickly and accurately positioned.
Because the selling of the cold area house resources needs to be pulled, the heat of the cold area house resources is improved, and in an optional embodiment, intention information of a real estate product supplier can be obtained; the property product corresponding to the intention information of the property product supplier is added to the target property product according to the intention information of the property product supplier. The method has the advantages that the house finding experience of the user is met, and meanwhile, some specific house sources are inserted into specific positions in the recommendation list of the target property product according to customized requirements of a merchant, such as a house-leaving inclination strategy, an operation activity strategy, advertisement promotion requirements and the like.
An embodiment of the present invention provides an electronic device, as shown in fig. 3, which includes one or more processors 31 and a memory 32, and one processor 33 is taken as an example in fig. 3.
The controller may further include: an input device 33 and an output device 34.
The processor 31, the memory 32, the input device 33 and the output device 34 may be connected by a bus or other means, and fig. 3 illustrates the connection by a bus as an example.
The processor 31 may be a Central Processing Unit (CPU). The processor 31 may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof. A general purpose processor may be a microprocessor or any conventional processor or the like.
The memory 32, which is a non-transitory computer readable storage medium, can be used for storing non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the control methods in the embodiments of the present application. The processor 31 executes various functional applications of the server and data processing, i.e., a property product searching method implementing the above-described method embodiments, by executing non-transitory software programs, instructions, and modules stored in the memory 32.
The memory 32 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of a processing device operated by the server, and the like. Further, the memory 32 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 32 may optionally include memory located remotely from the processor 31, which may be connected to a network connection device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 33 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the processing device of the server. The output device 34 may include a display device such as a display screen.
One or more modules are stored in the memory 32, which when executed by the one or more processors 31 perform the method as shown in fig. 1 or 2.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above may be implemented by instructing relevant hardware through a computer program, and the executed program may be stored in a computer-readable storage medium, and when executed, may include the processes of the embodiments of the property product searching method described above. The storage medium may be a magnetic disk, an optical disk, a Read-only memory (ROM), a Random Access Memory (RAM), a flash memory (FlashMemory), a hard disk (hard disk drive, abbreviated as HDD) or a Solid State Drive (SSD), etc.; the storage medium may also comprise a combination of memories of the kind described above.
Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope defined by the appended claims.
Claims (10)
1. A method for searching a property product, wherein the property product is preset with a state attribute with multi-level weight, the method comprises the following steps:
obtaining a search request, wherein the search request comprises a user intention of a multi-level weight;
matching the real estate product with the search request according to the user intention and the state attribute to obtain a matching result;
sequencing the matching results by utilizing a sequencing model to obtain a sequencing result, wherein the sequencing model is obtained by adopting multi-level weight training data training;
and obtaining the target property product according to the sequencing result.
2. The property product searching method of claim 1 wherein the search request includes multi-dimensional data information;
the matching the property product with the search request according to the user intent and the status attribute to obtain a matching result comprises:
respectively recalling the real estate products according to preset recalling rules corresponding to different dimensional data information based on the multi-dimensional data information, and establishing a plurality of recall index sets, wherein the recall index sets are respectively in one-to-one correspondence with the multi-dimensional data;
performing matching calculation on the real estate products and the search requests in the plurality of recall index sets in a hierarchical manner according to the weight levels of the user intention and the state attribute to obtain the matching degree;
performing preliminary sorting on the real estate products according to the matching degree to obtain a preliminary sorting result;
and taking the preliminary sorting result as the matching result.
3. The property product searching method of claim 2 wherein the multi-dimensional data information comprises: at least two of user demand information, user behavior information, user portrait information, and real estate product portrait information;
the recall index set includes: at least two of a search index collection, a behavior preference index collection, a user portrait index collection, a property product heat index collection.
4. The property product searching method of claim 3,
the method for establishing the search index set comprises the following steps:
recalling the user demand information by using a log analysis frame, and establishing a search index set;
the user behavior information includes: the method for establishing the behavior preference index set comprises the following steps:
recalling the behavior of the user based on the historical behavior information and the real-time behavior information of the user, and establishing a behavior preference index set;
the method for establishing the user portrait index set comprises the following steps:
acquiring a user portrait;
mining property product information corresponding to the user portrait in a preset property product database;
screening the real estate product information based on a location service, and establishing the user portrait index set;
the method for establishing the hot index set of the property product comprises the following steps:
acquiring a property product database corresponding to a search request based on the search request;
and screening the real estate product information with the heat degree larger than a preset value from the real estate product database corresponding to the search request, and establishing a real estate product heat degree index set.
5. The property product search method of claim 1 wherein the ranking the matching results using a ranking model resulting in ranked results comprises:
acquiring training data, wherein the training data comprises multi-level sample data and sample weight information, and the sample weight information comprises weight information of the user intention and weight information of the state attribute;
training a preset sequencing model by using the training data to obtain a trained sequencing model;
and sequencing the matching results by utilizing the trained sequencing model to obtain the sequencing result.
6. The property product searching method of claim 5 wherein the multiple levels of sample data comprise multiple levels of online sample data and multiple levels of offline sample data.
7. The property product searching method of claim 1, comprising, after obtaining the target property product according to the sorting result:
acquiring intention information of a supplier of a real estate product;
adding a property product corresponding to the intention information of the property product supplier to the target property product according to the intention information of the property product supplier.
8. The property product search method of claim 1 wherein the obtaining a search request comprises:
acquiring a search character string input by a user;
and processing at least one of recommendation, completion and consciousness recognition on the search character string to obtain the search request.
9. A computer-readable storage medium storing computer instructions for causing a computer to execute the property product searching method of any one of claims 1 to 8.
10. An electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to cause the at least one processor to perform the property product searching method of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911226750.XA CN110968801A (en) | 2019-12-04 | 2019-12-04 | Real estate product searching method, storage medium and electronic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911226750.XA CN110968801A (en) | 2019-12-04 | 2019-12-04 | Real estate product searching method, storage medium and electronic device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110968801A true CN110968801A (en) | 2020-04-07 |
Family
ID=70032970
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911226750.XA Pending CN110968801A (en) | 2019-12-04 | 2019-12-04 | Real estate product searching method, storage medium and electronic device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110968801A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111625680A (en) * | 2020-05-15 | 2020-09-04 | 青岛聚看云科技有限公司 | Method and device for determining search result |
| CN112487310A (en) * | 2020-10-23 | 2021-03-12 | 贝壳技术有限公司 | Cell recommendation method and system |
| CN115168436A (en) * | 2022-07-20 | 2022-10-11 | 贝壳找房(北京)科技有限公司 | Query information processing method, electronic device and readable storage medium |
| WO2022266973A1 (en) * | 2021-06-24 | 2022-12-29 | 曹庆恒 | Intelligent search method and system |
| CN115545853A (en) * | 2022-12-02 | 2022-12-30 | 云筑信息科技(成都)有限公司 | Searching method for searching suppliers |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104240163A (en) * | 2014-09-28 | 2014-12-24 | 百度在线网络技术（北京）有限公司 | Intelligent district recommendation method and intelligent district recommendation device |
| EP3057005A1 (en) * | 2015-02-12 | 2016-08-17 | Maurice Duault | Process for computing a score for a search engine used for accessing a database of real estate properties |
| CN108288179A (en) * | 2018-01-25 | 2018-07-17 | 链家网（北京）科技有限公司 | A kind of user preference source of houses computational methods and system |
| CN109190036A (en) * | 2018-08-24 | 2019-01-11 | 北京三快在线科技有限公司 | Recommended method, device, electronic equipment and storage medium |
| CN109377329A (en) * | 2018-12-25 | 2019-02-22 | 北京时光荏苒科技有限公司 | A kind of source of houses recommended method, device, storage medium and electronic equipment |
-
2019
- 2019-12-04 CN CN201911226750.XA patent/CN110968801A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104240163A (en) * | 2014-09-28 | 2014-12-24 | 百度在线网络技术（北京）有限公司 | Intelligent district recommendation method and intelligent district recommendation device |
| EP3057005A1 (en) * | 2015-02-12 | 2016-08-17 | Maurice Duault | Process for computing a score for a search engine used for accessing a database of real estate properties |
| CN108288179A (en) * | 2018-01-25 | 2018-07-17 | 链家网（北京）科技有限公司 | A kind of user preference source of houses computational methods and system |
| CN109190036A (en) * | 2018-08-24 | 2019-01-11 | 北京三快在线科技有限公司 | Recommended method, device, electronic equipment and storage medium |
| CN109377329A (en) * | 2018-12-25 | 2019-02-22 | 北京时光荏苒科技有限公司 | A kind of source of houses recommended method, device, storage medium and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| XUETAO GUO等: "intelligent e-government services with personalized recommendation techniques", 《INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS》 * |
| 夏正喜: "基于广义等级偏好优序法和层次分析法的购房决策", 《科技资讯》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111625680A (en) * | 2020-05-15 | 2020-09-04 | 青岛聚看云科技有限公司 | Method and device for determining search result |
| CN111625680B (en) * | 2020-05-15 | 2023-08-25 | 青岛聚看云科技有限公司 | Method and device for determining search results |
| CN112487310A (en) * | 2020-10-23 | 2021-03-12 | 贝壳技术有限公司 | Cell recommendation method and system |
| CN112487310B (en) * | 2020-10-23 | 2023-02-17 | 贝壳技术有限公司 | Cell recommendation method and system |
| WO2022266973A1 (en) * | 2021-06-24 | 2022-12-29 | 曹庆恒 | Intelligent search method and system |
| CN115168436A (en) * | 2022-07-20 | 2022-10-11 | 贝壳找房(北京)科技有限公司 | Query information processing method, electronic device and readable storage medium |
| CN115168436B (en) * | 2022-07-20 | 2023-08-08 | 贝壳找房(北京)科技有限公司 | Query information processing method, electronic device and readable storage medium |
| CN115545853A (en) * | 2022-12-02 | 2022-12-30 | 云筑信息科技(成都)有限公司 | Searching method for searching suppliers |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11580104B2 (en) | Method, apparatus, device, and storage medium for intention recommendation | |
| JP6708717B2 (en) | News recommendation method and device | |
| CN110968801A (en) | Real estate product searching method, storage medium and electronic device | |
| US9268826B2 (en) | System and method for crowdsourced template based search | |
| US8762383B2 (en) | Search engine and method for image searching | |
| CN110825957A (en) | Deep learning-based information recommendation method, device, equipment and storage medium | |
| CN110619076B (en) | Search term recommendation method and device, computer and storage medium | |
| CN110909182B (en) | Multimedia resource searching method, device, computer equipment and storage medium | |
| CN110825956A (en) | Information flow recommendation method and device, computer equipment and storage medium | |
| CN110472163B (en) | Method, device, electronic equipment and medium for determining display of map search result | |
| CN111814077B (en) | Information point query method, device, equipment and medium | |
| US10769136B2 (en) | Generalized linear mixed models for improving search | |
| CN111191133B (en) | Service search processing method, device and equipment | |
| CN114265981A (en) | Recommendation word determining method, device, equipment and storage medium | |
| CN105843817A (en) | Method and apparatus for searching on terminal device, and device | |
| CN112000763A (en) | Method, device, equipment and medium for determining competition relationship of interest points | |
| CN111143689A (en) | Method for constructing recommendation engine according to user requirements and user portrait | |
| CN111159559A (en) | Method for constructing recommendation engine according to user requirements and user behaviors | |
| CN113360788A (en) | Address recommendation method, device, equipment and storage medium | |
| CN114511085A (en) | Entity attribute value identification method, apparatus, device, medium, and program product | |
| CN110889029B (en) | Urban target recommendation method and device | |
| KR101747532B1 (en) | Method and system for recommending course for travel related query | |
| CN110929169A (en) | Position recommendation method based on improved Canopy clustering collaborative filtering algorithm | |
| CN111831892A (en) | Information recommendation method, information recommendation device, server and storage medium | |
| CN114237588A (en) | Code warehouse selection method, device, equipment and storage medium |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200407 |