CN116108269A - Search result display method and device, electronic equipment and storage medium - Google Patents
Search result display method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN116108269A CN116108269A CN202211733384.9A CN202211733384A CN116108269A CN 116108269 A CN116108269 A CN 116108269A CN 202211733384 A CN202211733384 A CN 202211733384A CN 116108269 A CN116108269 A CN 116108269A
- Authority
- CN
- China
- Prior art keywords
- search
- user
- sample
- search result
- prediction model
- 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
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The embodiment of the invention provides a search result display method, a search result display device, electronic equipment and a storage medium. The method comprises the following steps: acquiring a plurality of search results corresponding to target search words returned by a server; obtaining user interest degrees corresponding to the plurality of search results through a local pre-trained user interest degree prediction model; the training sample of the user interest degree prediction model at least comprises: sample user behavior characteristics of search words; the sample search term includes: search terms input by the user in history, and associative search terms of the search terms input by the user in history; rearranging the plurality of search results according to the sequence of the user interest level from high to low; and displaying the rearranged multiple search results. By the method, the user preference can be analyzed in a new dimension, so that the model effect of the user interest degree prediction model trained according to the behavior data of the user aiming at the search word is better, and the rearrangement result is more prone to user preference.
Description
Technical Field
The present invention relates to the field of terminal intelligence technologies, and in particular, to a search result display method, a search result display device, an electronic device, and a storage medium.
Background
With the continuous development of internet technology, terminal intelligence is a technology which is gradually rising and widely applied. At present, user preference scoring can be performed on search results searched by a user through a model at an end side, and the search results are rearranged according to user preference, so that the user can easily see the search results with preference.
At present, the models for scoring the user preference aiming at the search results are trained through training samples related to the search results, but the prediction effect of the user scoring model trained by the method is poor, the estimation accuracy of the search results is low, and the preference of the user cannot be predicted accurately.
Disclosure of Invention
The embodiment of the invention provides a search result display method, a device, electronic equipment and a storage medium, wherein user preference is analyzed by new dimension, namely, a user interest degree prediction model is trained by behavior data of a user aiming at search words, so that the estimation accuracy of the model on search results is improved.
The first aspect of the embodiment of the invention provides a search result display method, which is applied to a client, and comprises the following steps:
Acquiring a plurality of search results corresponding to target search words returned by a server;
obtaining user interest degrees corresponding to the plurality of search results through a local pre-trained user interest degree prediction model; the training sample of the user interest degree prediction model at least comprises: sample user behavior characteristics of search words; the sample search term includes: search terms input by the user in history, and associative search terms of the search terms input by the user in history;
rearranging the plurality of search results according to the sequence of the user interest level from high to low;
and displaying the rearranged multiple search results.
Optionally, the training sample of the user interest degree prediction model further includes: user behavior characteristics of sample search results; the sample search results are search results which are returned by the server and correspond to target sample search words, and the target sample search words are: the search term input by the user history, or the associated search term of the search term input by the user history.
Optionally, the method further comprises:
acquiring a plurality of associative search terms corresponding to search terms input by a user;
obtaining user interest degrees corresponding to the plurality of associative search words through the user interest degree prediction model;
Rearranging the plurality of associative search words according to the sequence of the user interest level from high to low;
displaying the rearranged plurality of associative search words.
Optionally, the method further comprises:
in the event that a user operation on the sample search term is detected, storing a positive training sample, the positive training sample comprising at least one of the following data: the occurrence time of the operation, the characteristics of the sample search word and the service type corresponding to the sample search word;
in the event that no user manipulation of the sample search term is detected, storing a negative training sample comprising at least one of the following data: and displaying the occurrence time of the sample search word, the characteristics of the sample search word and the service type corresponding to the sample search word.
Optionally, the method further comprises:
in the event that user manipulation of the sample search results is detected, storing a positive training sample comprising at least one of the following data: the occurrence time of the clicking operation, the label characteristics corresponding to the sample search result, the browsing duration of the sample search result and the service type corresponding to the sample search result;
In the event that no user manipulation of the sample search results is detected, storing a negative training sample comprising at least one of the following data: and displaying the occurrence time of the sample search result, the label characteristic corresponding to the sample search result, the browsing duration of zero of the sample search result and the service type corresponding to the sample search result.
Optionally, the method further comprises:
pulling the stored training samples in response to a user start operation;
updating model parameters of the user interest degree prediction model by using the stored training samples;
the obtaining the user interest degree corresponding to the plurality of search results through the local pre-trained user interest degree prediction model comprises the following steps:
and obtaining the user interest degree corresponding to the plurality of search results through the updated user interest degree prediction model.
A second aspect of an embodiment of the present invention provides a search result display apparatus, applied to a client, where the apparatus includes:
the first acquisition module is used for acquiring a plurality of search results corresponding to the target search word returned by the server;
The first prediction module is used for obtaining the user interestingness corresponding to the plurality of search results through a local pre-trained user interestingness prediction model; the training sample of the user interest degree prediction model at least comprises: sample user behavior characteristics of search words; the sample search term includes: search terms input by the user in history, and associative search terms of the search terms input by the user in history;
the first rearrangement module is used for rearranging the plurality of search results according to the sequence from high to low of the interest degree of the user;
and the first display module is used for displaying the rearranged plurality of search results.
Optionally, the training sample of the user interest degree prediction model further includes: user behavior characteristics of sample search results; the sample search results are search results which are returned by the server and correspond to target sample search words, and the target sample search words are: the search term input by the user history, or the associated search term of the search term input by the user history.
Optionally, the apparatus further includes:
a second acquisition module for acquiring a plurality of associative search terms corresponding to the search term input by the user;
The second prediction module is used for obtaining the user interest degree corresponding to the plurality of associative search words through the user interest degree prediction model;
the second rearrangement module is used for rearranging the plurality of associative search words according to the sequence from high to low of the interest degree of the user;
and the second display module is used for displaying the rearranged multiple associative search words.
Optionally, the apparatus further includes:
a first storage module, configured to store a positive training sample in a case that an operation of the sample search word by a user is detected, where the positive training sample includes at least one of the following data: the occurrence time of the operation, the characteristics of the sample search word and the service type corresponding to the sample search word;
and the second storage module is used for storing a negative training sample in the condition that the operation of the user on the sample search word is not detected, wherein the negative training sample comprises at least one of the following data: and displaying the occurrence time of the sample search word, the characteristics of the sample search word and the service type corresponding to the sample search word.
Optionally, the apparatus further includes:
a third storage module, configured to store a positive training sample in a case where an operation of the sample search result by the user is detected, where the positive training sample includes at least one of the following data: the occurrence time of the clicking operation, the label characteristics corresponding to the sample search result, the browsing duration of the sample search result and the service type corresponding to the sample search result;
A fourth storage module, configured to store a negative training sample in a case where no operation of the sample search result by the user is detected, where the negative training sample includes at least one of the following data: and displaying the occurrence time of the sample search result, the label characteristic corresponding to the sample search result, the browsing duration of zero of the sample search result and the service type corresponding to the sample search result.
Optionally, the apparatus further includes:
the third acquisition module is used for responding to the starting operation of a user and pulling the stored training samples;
the model updating module is used for updating the model parameters of the user interest degree prediction model by utilizing the stored training samples;
the first prediction module includes:
and the first prediction sub-module is used for obtaining the user interestingness corresponding to the plurality of search results through the updated user interestingness prediction model.
A third aspect of the embodiments of the present invention provides an electronic device, including a processor, a memory, and a computer program stored on the memory and executable on the processor, the computer program implementing the steps of the search result presentation method according to the first aspect of the present invention when executed by the processor.
A fourth aspect of the embodiments of the present invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the search result presentation method according to the first aspect of the present invention.
According to the embodiment of the invention, the client acquires a plurality of search results which are returned by the server and are aimed at target search words, and the user interest degree corresponding to the plurality of search results is predicted and obtained through a user interest degree prediction model which is trained in advance locally by the client, so that the plurality of search results are rearranged and displayed according to the sequence of the user interest degree from high to low; the training samples of the user interest degree prediction model in this embodiment at least include: sample user behavioral characteristics of the search term. According to the search result display method, the client analyzes user preferences through a new dimension, namely, the behavior data of the user aiming at the search words can better reflect the actual preferences of the user, so that the client performs model training in advance according to the collected user behavior characteristics of the user aiming at the sample search words (the search words input by the user history and the associated search words of the search words input by the user history), a trained user interest degree prediction model is obtained, the user interest degree prediction model estimates and predicts the search results more accurately and is more biased to the actual preferences of the user, the prediction effect of the model is better improved, when the client obtains the search results returned by the server, the trained user interest degree prediction model predicts a plurality of search results, rearranges and displays the search results according to the predicted user interest degree, and the rearranged results are more prone to the preferences of the user, so that the user can find the preference targets of the user more easily.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, 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 flow chart of a search result presentation method according to an embodiment of the present invention;
FIG. 2 is a block diagram of a search result presentation apparatus according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, fig. 1 is a flowchart of a search result display method according to an embodiment of the invention. The model running method provided by the embodiment can be applied to the client. In this embodiment, the client may refer to a program that provides a local service for a user, and the client may implement the same software function in different program forms; for example, the client may be an application software APP, applet, web browser, or the like. As shown in fig. 1, the search result display method of the present embodiment may include the following steps:
Step S11: and acquiring a plurality of search results corresponding to the target search word returned by the server.
In this embodiment, when the user uses the search function of the client, the client may send a search request to the server for the target search word, so as to request a plurality of search results corresponding to the target search word to the server. The target search word is a word which is determined by a user and needs to perform a search function. After receiving the search request, the server determines a plurality of search results corresponding to the target search word according to the target search word carried by the search request and returns the search results to the client, and the client acquires the plurality of search results corresponding to the target search word returned by the server.
Step S12: and obtaining the user interestingness corresponding to the plurality of search results through a local pre-trained user interestingness prediction model.
In this embodiment, the client is pre-trained with a user interest level prediction model, and the user interest level prediction model is used for predicting the user interest level of the user for the search result. In this embodiment, the training samples of the client training user interest degree prediction model at least include user behavior features of sample search words, where the sample search words include: search terms historically entered by the user, and associative search terms for the search terms historically entered by the user.
Wherein, the associative search term is a related and similar term having an association relationship with the search term. For example, the search term is "potato," then the associative search term may be: potatoes, potato braised meats, potato nets, and the like, related, similar terms to the search term "potato" in association with each other. The associated search word may be determined by the client directly according to the search word, or may be determined by the server according to the search word and returned to the client, and the client and the server may determine the associated search word of each search word by any available method, which is not limited in this embodiment of the present application. For example, the plurality of association words and the degree of correlation of each search word may be determined through a trained association word prediction model, or through other prediction methods, and TopN or an association word with a degree of correlation exceeding a certain threshold may be selected from the plurality of association words and the degree of correlation as the association search word of the corresponding search word, where N may be a positive integer.
The search words input by the user in the history are the history search words input by the user in the client acquired by the client, the associated search words of the search words input by the user in the history are related and similar words having association relation with the search words input by the user in the history, the associated search words of the search words input by the user in the history can be determined and returned by the server according to the history search words input by the user in the client, or the client can determine the associated search words according to the history search words input by the user in the client.
And the client inputs a plurality of search results corresponding to the target search word returned by the server into the user interest degree prediction model, performs model prediction, and obtains the user interest degree corresponding to each of the plurality of search results output by the user interest degree prediction model.
In this embodiment, when the user uses the search function of the client, after the user inputs the search word, the client further determines and displays a plurality of association words of the search word, so that the user can determine the target search word from the input search word and the plurality of association words of the search word to search. In the process of determining target search words, a client records the user behavior characteristics aiming at the search words, and the user behavior characteristics of the search words recorded each time can be used as the user behavior characteristics of the sample search words to train a user interestingness prediction model. In this embodiment, the user interest degree prediction model is trained through the client, and since data collection (i.e., collection of training samples) and use are both performed at the client, the real-time performance is better provided and privacy leakage can be avoided, meanwhile, all data come from the user individual, and the prediction result is more prone to preference of a single user, so that the prediction accuracy of the model is further improved.
Step S13: and rearranging the plurality of search results according to the sequence of the user interest level from high to low.
In this embodiment, it may be understood that the plurality of search results returned by the server are arranged in sequence, and after obtaining the user interest level corresponding to the plurality of search results through the locally stored user interest level prediction model, the client may rearrange the plurality of search results according to the order of the user interest level from high to low.
Step S14: and displaying the rearranged multiple search results.
In this embodiment, after the client rearranges the plurality of search results, the rearranged plurality of search results may be displayed in the client, so that the user may more easily see the target search result of the user, and user experience may be improved.
In this embodiment, the client performs model training in advance according to the collected user behavior characteristics of the sample search word (the search word input by the user in history and the associated search word of the search word input by the user in history) to obtain a trained user interest prediction model. The client analyzes the user preference by a new dimension, the new dimension (the behavior data of the user aiming at the search word) can better reflect the actual preference of the user, so that the user interest degree prediction model trained by the client predicts the search result more accurately and is more biased to the actual preference of the user, the prediction effect of the model is better improved, when the client acquires the search result returned by the server, a plurality of search results can be predicted by the trained user interest degree prediction model, and the search results are rearranged and displayed according to the predicted user interest degree, so that the rearranged result is more prone to the user preference, the user can find the preference target more easily, and the user experience is improved.
In combination with the above embodiment, in an implementation manner, the embodiment of the present invention further provides a search result display method. In the method, the training sample of the user interest degree prediction model further comprises: user behavior characteristics of sample search results; the sample search result is a search result returned by the server and corresponds to a target sample search word, wherein the target sample search word is one search word determined by a user among sample search words (search words input by the user in history and associated search words of the search words input by the user in history) so as to search for the determined search word. The target sample search term is: the search term input by the user history, or the associated search term of the search term input by the user history.
In this embodiment, in order to further improve the prediction accuracy of the user interest degree prediction model, avoid analyzing too many sides of data preferred by the user, the training sample used for training the user interest degree prediction model by the client includes, in addition to the user behavior feature of the sample search word, the user behavior feature of the sample search result.
The method comprises the steps that a sample search result is a search result which is returned by a server side aiming at a target sample search word and corresponds to the target sample search word, the target sample search word is a target search word collected by a client side in a history, and the target sample search word is: search terms that the user has historically entered, or associative search terms that the user has historically entered. That is, the sample search results are the search results corresponding to the target search word returned by the server and obtained by the client history.
In this embodiment, when the user uses the search function of the client, after the client displays the search result, the user performs related operations on the search result, such as browsing or clicking on the search result. At this time, the client records the user behavior characteristics aiming at the search results, and the user behavior characteristics of the search results recorded each time can be used as the user behavior characteristics of the sample search results to train the user interestingness prediction model.
In this embodiment, the client trains the user interest degree prediction model through the user behavior feature of the sample search word and the user behavior feature of the sample search result together, so as to perform correlation analysis on the user preference in multiple directions and at various angles, and avoid analyzing data of the user preference from being too onesided, thereby further improving the prediction accuracy of the user interest degree prediction model and playing a better effect on rearrangement of the search result.
In combination with the above embodiment, in an implementation manner, the embodiment of the present invention further provides a search result display method. In the method, in addition to the above steps, the method may further include step S21:
step S21: a plurality of associative search terms corresponding to search terms entered by a user are obtained.
In this embodiment, when the user uses the search function of the client, the user may input a search term in the client, and after the client determines the search term input by the user, the client may send a search term request to the server for the search term input by the user, so as to request a plurality of associated search terms corresponding to the search term input by the user to the server. After receiving the search word request, the server determines a plurality of associated search words corresponding to the search words input by the user according to the search words input by the user carried by the search word request, and returns the search words to the client, and the client acquires the plurality of associated search words corresponding to the search words input by the user returned by the server.
Step S22: and obtaining the user interest degree corresponding to the plurality of associative search words through the user interest degree prediction model.
In this embodiment, the client locally pre-trains the user interest level prediction model to estimate the user interest level of the association search term. The client inputs a plurality of association search words corresponding to the search words input by the user and returned by the server to the user interest degree prediction model, and performs model prediction to obtain the user interest degree corresponding to the association search words output by the user interest degree prediction model.
Step S23: and rearranging the plurality of associative search words according to the sequence of the user interest level from high to low.
In this embodiment, it may be understood that the plurality of association search words returned by the server side are arranged in sequence, and after the client side obtains the user interest degrees corresponding to the plurality of association search words through the locally stored user interest degree prediction model, the plurality of association search words may be rearranged according to the order of the user interest degrees from high to low.
Step S24: displaying the rearranged plurality of associative search words.
In this embodiment, after the client rearranges the plurality of associative search words, the rearranged plurality of associative search words may be displayed in the client.
In this embodiment, the user interest prediction model trained in the client may also perform estimation and rearrangement of the user interest by using multiple associated search words returned by the server, so that the user may see the search words input by the user and multiple associated search words reordered according to the user preference at the same time in the client, so that the user may find the target search word that the user wants to search more easily, and the user may find his preference target more easily, thereby improving user experience and promoting development of search service.
In combination with the above embodiment, in an implementation manner, the embodiment of the present invention further provides a search result display method. In the method, in addition to the above steps, the method may further include steps S31 and S32:
step S31: in the event that a user operation on the sample search term is detected, storing a positive training sample, the positive training sample comprising at least one of the following data: the occurrence time of the operation, the characteristics of the sample search word and the service type corresponding to the sample search word.
In this embodiment, when the client detects that the user performs an operation on a certain sample search word during a display period of the sample search word after the user inputs the search word once, determining a user behavior feature corresponding to the sample search word, and recording and storing the user behavior feature corresponding to the sample search word as a positive training sample, where the positive training sample includes at least one of the following data: the user aims at the occurrence time of the operation of the sample search word and the characteristics of the sample search word, and the service type corresponding to the sample search word.
The above-mentioned operation may be a click or confirmation operation of the user with respect to a sample search word (a search word input by the user in history, and an associated search word of a search word input by the user in history), wherein the confirmation operation may be a voice control confirmation operation or the like. The service type corresponding to the sample search word in this embodiment may refer to the classification represented by the sample search word.
Step S32: in the event that no user manipulation of the sample search term is detected, storing a negative training sample comprising at least one of the following data: and displaying the occurrence time of the sample search word, the characteristics of the sample search word and the service type corresponding to the sample search word.
In this embodiment, when the operation of the user on one or more sample search words is not detected during the presentation of the sample search words after the user inputs the search word once, the client determines, for each sample search word that is not subjected to the operation by the user, a user behavior feature corresponding to each sample search word by the user, and records and stores the user behavior feature corresponding to each sample search word as a negative training sample, where the negative training sample includes at least one of the following data: the client displays the occurrence time of the sample search word, the characteristics of the sample search word and the service type corresponding to the sample search word.
In this embodiment, the client stores the user behavior feature of the operation of the user with respect to the sample search word as a positive training sample each time, and stores the user behavior feature of the operation of the user with respect to the sample search word as a negative training sample each time the client does not detect the operation of the user with respect to the sample search word, so as to continuously collect the user behavior data of the user with respect to the sample search word occurring in the client, so as to train the training sample of the user interest degree prediction model in the client, thereby training the model more biased to the user preference.
In combination with the above embodiment, in an implementation manner, the embodiment of the present invention further provides a search result display method. In the method, in addition to the above steps, the method may further include steps S41 and S42:
step S41: in the event that user manipulation of the sample search results is detected, storing a positive training sample comprising at least one of the following data: the occurrence time of the clicking operation, the label characteristics corresponding to the sample search result, the browsing duration of the sample search result and the service type corresponding to the sample search result.
In this embodiment, when detecting an operation of a user on one or more sample search results during a presentation period of a plurality of sample search results searched at a time, the client determines a user behavior feature corresponding to the sample search result, and records and stores the user behavior feature corresponding to the sample search result as a positive training sample, where the positive training sample includes at least one of the following data: the method comprises the steps of generating time when a user clicks a sample search result, label characteristics corresponding to the sample search result, browsing duration of the sample search result and service types corresponding to the sample search result.
The browsing time of the sample search result in this embodiment is the time from the client to the return of the details page of the sample search result; the type of service to which the sample search results correspond may refer to the classification represented by the tag of the sample search results.
Step S42: in the event that no user manipulation of the sample search results is detected, storing a negative training sample comprising at least one of the following data: and displaying the occurrence time of the sample search result, the label characteristic corresponding to the sample search result, the browsing duration of zero of the sample search result and the service type corresponding to the sample search result.
In this embodiment, when the client does not detect an operation of the user on one or more sample search results during a presentation period of a plurality of sample search results searched at a time, the client determines, for each sample search result that is not clicked by the user, a user behavior feature corresponding to each sample search result by the user, and records and stores the user behavior feature corresponding to each sample search result as a negative training sample, where the negative training sample includes at least one of the following data: the client displays the occurrence time of the sample search result, the label characteristics corresponding to the sample search result, the browsing duration of zero sample search result and the service type corresponding to the sample search result.
In this embodiment, the client stores the user behavior feature of each time the user detects the click operation of the user on the sample search result as a positive training sample, and stores the user behavior feature of each time the user does not detect the click operation of the user on the sample search result as a negative training sample, so as to continuously collect the user behavior data of the user on the sample search result generated in the client, and train the training sample of the user interest degree prediction model in the client, so as to train the model more biased to the user preference.
In one embodiment, training samples of the predictive model at the user's interest level include: when the user behavior feature of the sample search term and the user behavior feature of the sample search result are the same, one negative training sample in the training samples may include at least one of the following data: the client displays the occurrence time of the sample search word and the characteristics of the sample search word, and the service type corresponding to the sample search word; alternatively, a negative training sample may include at least one of the following data: the client displays the occurrence time of the sample search result, the label characteristics corresponding to the sample search result, the browsing duration of zero sample search result and the service type corresponding to the sample search result.
One of the training samples may include at least one of the following data: the user aims at the occurrence time of the operation of the sample search word and the characteristics of the sample search word, and the service type corresponding to the sample search word; alternatively, a positive training sample may include at least one of the following data: the method comprises the steps of generating time when a user clicks a sample search result, label characteristics corresponding to the sample search result, browsing duration of the sample search result and service types corresponding to the sample search result.
In one embodiment, the initial model of the user interest level prediction model may be a linear regression, and the input data of the user interest level prediction model is: the current time, the characteristics of the search word/the label characteristics corresponding to the search result, and the service type corresponding to the search result/the service type corresponding to the search word; the output data of the user interest degree prediction model is browsing time possibly generated by the user. Since the sample search word used for training has no browsing time, the weight of the browsing time of the sample search word in the positive and negative training samples of the sample search word can be set to be the lowest during training, and the influence of the browsing time of the sample search word on the model can be reduced to the lowest, for example, the weight of the browsing time of the search word can be set to be 0 to be used as calibration.
In combination with any of the above embodiments, in an implementation manner, the embodiment of the present invention further provides a search result display method. In the method, the method may further include steps S51 and S52 in addition to the above steps, and the above step S12 may specifically include step S53:
step S51: and pulling the stored training samples in response to a user start operation.
In this embodiment, in response to a user's start operation, the client pulls the historically stored training samples from the local store, where the training samples include a positive training sample and a negative training sample.
Step S52: and updating the model parameters of the user interest degree prediction model by using the stored training samples.
In this embodiment, after the client pulls the training samples stored in the local history, a new training sample in the training samples stored in the history may be used as training data, and the model parameters of the latest user interest prediction model stored in the local may be updated to obtain an updated user interest prediction model, where the updated user interest prediction model may perform search results or estimation of the user interest of the search term.
Step S53: and obtaining the user interest degree corresponding to the plurality of search results through the updated user interest degree prediction model.
In this embodiment, the client inputs a plurality of search results corresponding to the target search word returned by the server to the updated user interest degree prediction model, and performs model prediction to obtain the user interest degree corresponding to each of the plurality of search results output by the updated user interest degree prediction model.
In this embodiment, the client may update the user interest prediction model according to the user behavior data of each user aiming at the search word or the search result in the client, so that the prediction result is more prone to personal habits and hobbies of the user, thereby improving user experience and promoting business development.
In one embodiment, after the client boots up, the client may pull all of the historically stored training samples from the local store, including positive training samples and negative training samples, when it is detected in the client that the user interest prediction model is not locally saved. After the client pulls all training samples stored in the local history, the client can use all the training samples stored in the history as training data to perform model training so as to train to obtain a user interest degree prediction model, and the user interest degree prediction model is stored in a storage space of the client.
In an embodiment, the method for displaying the search result may include the following steps:
1. when receiving a search word input by a user, the client can determine a plurality of association words of the search word, wherein the association words are words related to and similar to the search word input by the user, and the association words are displayed in a list form in a client interface;
2. the user directly clicks a confirmation key beside the search bar to determine the search word input by the user, or clicks the association word to search the clicked association word to determine the content (namely the target search word) which the user wants to search, and then the client interacts with the server to receive the search result corresponding to the target search word returned by the server and display the search result in a list form;
3. the user can browse the content of the search result, click on the content which is wanted to be seen, enter the corresponding detail page after clicking, and click on the content which is not wanted to be seen.
4. The client takes the user behavior data corresponding to the search word input when the user clicks the confirmation key each time and the user behavior data corresponding to the associated word clicked each time by the user as a positive sample, wherein the positive sample data comprises but is not limited to: clicking time and content of words, and corresponding business types of the words; in addition, the client may take, as a negative sample, user behavior data of the associational words that have been shown but not clicked or the input search words when the user selects the search words, where the negative sample data includes, but is not limited to: and displaying the occurrence time and the content of the words, wherein the words correspond to the service types.
5. Click behavior data of a user for search results presented by the client is used by the client as a positive sample, including but not limited to: the occurrence time of clicking the search result, the label word of the search result, the browsing time of the search result, the service type of accessing the search result and the like; in addition, after the whole search function is used, the data corresponding to other search results which are not clicked are used as a piece of negative sample data, and the negative sample data includes but is not limited to: the time when the search result presentation occurs, the tag term of the search result, the browsing duration of the search result is 0, the service type of the unaccessed search result, etc.
6. After each start, the client takes all the positive and negative samples stored locally as training data to perform model training, and stores the trained model in a local storage space. And if the model is already stored in the local area, reading the model and updating the model according to the new training sample. The model is a linear regression, inputs as the current time, the tag feature of the associated word feature or the search result, the service type (the service type corresponding to the word or the search result), and outputs as the browsing time which the user may generate. And the associated word sample has no browsing time, so the weight of the browsing time of the search word sample data can be set to be the lowest during training, the influence on the model is minimized, and the weight of the browsing time of the search word sample data is set to be 0 as calibration.
7. So far, when the user uses the search function at the client, and the client displays the associated words, the client can predict each associated word in the list data of the associated words through the model, rearrange the corresponding associated word data according to the obtained prediction result in a big-to-small arrangement mode, and the rearranged associated word list is the arrangement according to the user interested value from the data.
8. Similarly, when the client displays the search results, each search result in the list data of the search results can be predicted through the model, the corresponding search result data is rearranged according to the obtained prediction results in a big-to-small arrangement mode, and the rearranged search result list is the arrangement of the values of interest of the user on the data.
9. The model of the client is updated every time the user searches and browses later, so that the predicted result is more prone to personal habit and hobbies of the user, the user experience is improved, and the business development is promoted.
The server in this embodiment may be understood as a server, which may be an independent server, or may be a server cluster or a distributed system of multiple physical server projects, and may also be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network content delivery network), and basic cloud computing services such as big data and an artificial intelligence platform.
It should be noted that, for simplicity of description, the method embodiments are shown as a series of acts, but it should be understood by those skilled in the art that the embodiments are not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the embodiments. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred embodiments, and that the acts are not necessarily required by the embodiments of the invention.
Based on the same inventive concept, an embodiment of the present invention provides a search result presentation apparatus 200, and the search result presentation apparatus 200 is applicable to a client. Referring to fig. 2, fig. 2 is a block diagram illustrating a search result display apparatus according to an embodiment of the present invention. As shown in fig. 2, the search result presentation apparatus 200 includes:
a first obtaining module 201, configured to obtain a plurality of search results corresponding to the target search term returned by the server;
a first prediction module 202, configured to obtain user interestingness corresponding to the plurality of search results through a local pre-trained user interestingness prediction model; the training sample of the user interest degree prediction model at least comprises: sample user behavior characteristics of search words; the sample search term includes: search terms input by the user in history, and associative search terms of the search terms input by the user in history;
A first rearrangement module 203, configured to rearrange the plurality of search results in order of from high to low interest level of the user;
the first display module 204 is configured to display the rearranged plurality of search results.
Optionally, the training sample of the user interest degree prediction model further includes: user behavior characteristics of sample search results; the sample search results are search results which are returned by the server and correspond to target sample search words, and the target sample search words are: the search term input by the user history, or the associated search term of the search term input by the user history.
Optionally, the apparatus 200 further includes:
a second acquisition module for acquiring a plurality of associative search terms corresponding to the search term input by the user;
the second prediction module is used for obtaining the user interest degree corresponding to the plurality of associative search words through the user interest degree prediction model;
the second rearrangement module is used for rearranging the plurality of associative search words according to the sequence from high to low of the interest degree of the user;
and the second display module is used for displaying the rearranged multiple associative search words.
Optionally, the apparatus 200 further includes:
A first storage module, configured to store a positive training sample in a case that an operation of the sample search word by a user is detected, where the positive training sample includes at least one of the following data: the occurrence time of the operation, the characteristics of the sample search word and the service type corresponding to the sample search word;
and the second storage module is used for storing a negative training sample in the condition that the operation of the user on the sample search word is not detected, wherein the negative training sample comprises at least one of the following data: and displaying the occurrence time of the sample search word, the characteristics of the sample search word and the service type corresponding to the sample search word.
Optionally, the apparatus 200 further includes:
a third storage module, configured to store a positive training sample in a case where an operation of the sample search result by the user is detected, where the positive training sample includes at least one of the following data: the occurrence time of the clicking operation, the label characteristics corresponding to the sample search result, the browsing duration of the sample search result and the service type corresponding to the sample search result;
a fourth storage module, configured to store a negative training sample in a case where no operation of the sample search result by the user is detected, where the negative training sample includes at least one of the following data: and displaying the occurrence time of the sample search result, the label characteristic corresponding to the sample search result, the browsing duration of zero of the sample search result and the service type corresponding to the sample search result.
Optionally, the apparatus 200 further includes:
the third acquisition module is used for responding to the starting operation of a user and pulling the stored training samples;
the model updating module is used for updating the model parameters of the user interest degree prediction model by utilizing the stored training samples;
the first prediction module 202 includes:
and the first prediction sub-module is used for obtaining the user interestingness corresponding to the plurality of search results through the updated user interestingness prediction model.
Based on the same inventive concept, another embodiment of the present invention provides an electronic device 300, as shown in fig. 3. Fig. 3 is a schematic diagram of an electronic device according to an embodiment of the present invention. The electronic device comprises a processor 301, a memory 302 and a computer program stored on the memory 302 and executable on the processor 301, which when executed by the processor implements the steps of the search result presentation method according to any of the embodiments of the present invention. By way of example, the client in any of the embodiments described above may be a computer program running on an electronic device.
Based on the same inventive concept, another embodiment of the present invention provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the search result presentation method according to any of the above embodiments of the present invention.
In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described by differences from other embodiments, and identical and similar parts between the embodiments are all enough to be referred to each other.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising instructions for causing a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those having ordinary skill in the art without departing from the spirit of the present invention and the scope of the claims, which are to be protected by the present invention.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk, etc.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
Claims (14)
1. A search result presentation method, applied to a client, the method comprising:
acquiring a plurality of search results corresponding to target search words returned by a server;
obtaining user interest degrees corresponding to the plurality of search results through a local pre-trained user interest degree prediction model; the training sample of the user interest degree prediction model at least comprises: sample user behavior characteristics of search words; the sample search term includes: search terms input by the user in history, and associative search terms of the search terms input by the user in history;
rearranging the plurality of search results according to the sequence of the user interest level from high to low;
and displaying the rearranged multiple search results.
2. The search result presentation method of claim 1, wherein the training sample of the user interest prediction model further comprises: user behavior characteristics of sample search results; the sample search results are search results which are returned by the server and correspond to target sample search words, and the target sample search words are: the search term input by the user history, or the associated search term of the search term input by the user history.
3. The search result presentation method of claim 1, further comprising:
acquiring a plurality of associative search terms corresponding to search terms input by a user;
obtaining user interest degrees corresponding to the plurality of associative search words through the user interest degree prediction model;
rearranging the plurality of associative search words according to the sequence of the user interest level from high to low;
displaying the rearranged plurality of associative search words.
4. The search result presentation method of claim 1, further comprising:
in the event that a user operation on the sample search term is detected, storing a positive training sample, the positive training sample comprising at least one of the following data: the occurrence time of the operation, the characteristics of the sample search word and the service type corresponding to the sample search word;
in the event that no user manipulation of the sample search term is detected, storing a negative training sample comprising at least one of the following data: and displaying the occurrence time of the sample search word, the characteristics of the sample search word and the service type corresponding to the sample search word.
5. The search result presentation method of claim 2, further comprising:
in the event that user manipulation of the sample search results is detected, storing a positive training sample comprising at least one of the following data: the occurrence time of the clicking operation, the label characteristics corresponding to the sample search result, the browsing duration of the sample search result and the service type corresponding to the sample search result;
in the event that no user manipulation of the sample search results is detected, storing a negative training sample comprising at least one of the following data: and displaying the occurrence time of the sample search result, the label characteristic corresponding to the sample search result, the browsing duration of zero of the sample search result and the service type corresponding to the sample search result.
6. The method of any one of claims 1 to 5, further comprising:
pulling the stored training samples in response to a user start operation;
updating model parameters of the user interest degree prediction model by using the stored training samples;
The obtaining the user interest degree corresponding to the plurality of search results through the local pre-trained user interest degree prediction model comprises the following steps:
and obtaining the user interest degree corresponding to the plurality of search results through the updated user interest degree prediction model.
7. A search result presentation apparatus, for application to a client, the apparatus comprising:
the first acquisition module is used for acquiring a plurality of search results corresponding to the target search word returned by the server;
the first prediction module is used for obtaining the user interestingness corresponding to the plurality of search results through a local pre-trained user interestingness prediction model; the training sample of the user interest degree prediction model at least comprises: sample user behavior characteristics of search words; the sample search term includes: search terms input by the user in history, and associative search terms of the search terms input by the user in history;
the first rearrangement module is used for rearranging the plurality of search results according to the sequence from high to low of the interest degree of the user;
and the first display module is used for displaying the rearranged plurality of search results.
8. The search result presentation apparatus of claim 7, wherein the training sample of the user interest prediction model further comprises: user behavior characteristics of sample search results; the sample search results are search results which are returned by the server and correspond to target sample search words, and the target sample search words are: the search term input by the user history, or the associated search term of the search term input by the user history.
9. The search result presentation apparatus of claim 7, wherein the apparatus further comprises:
a second acquisition module for acquiring a plurality of associative search terms corresponding to the search term input by the user;
the second prediction module is used for obtaining the user interest degree corresponding to the plurality of associative search words through the user interest degree prediction model;
the second rearrangement module is used for rearranging the plurality of associative search words according to the sequence from high to low of the interest degree of the user;
and the second display module is used for displaying the rearranged multiple associative search words.
10. The search result presentation apparatus of claim 7, wherein the apparatus further comprises:
a first storage module, configured to store a positive training sample in a case that an operation of the sample search word by a user is detected, where the positive training sample includes at least one of the following data: the occurrence time of the operation, the characteristics of the sample search word and the service type corresponding to the sample search word;
and the second storage module is used for storing a negative training sample in the condition that the operation of the user on the sample search word is not detected, wherein the negative training sample comprises at least one of the following data: and displaying the occurrence time of the sample search word, the characteristics of the sample search word and the service type corresponding to the sample search word.
11. The search result presentation apparatus of claim 8, wherein the apparatus further comprises:
a third storage module, configured to store a positive training sample in a case where an operation of the sample search result by the user is detected, where the positive training sample includes at least one of the following data: the occurrence time of the clicking operation, the label characteristics corresponding to the sample search result, the browsing duration of the sample search result and the service type corresponding to the sample search result;
a fourth storage module, configured to store a negative training sample in a case where no operation of the sample search result by the user is detected, where the negative training sample includes at least one of the following data: and displaying the occurrence time of the sample search result, the label characteristic corresponding to the sample search result, the browsing duration of zero of the sample search result and the service type corresponding to the sample search result.
12. The search result presentation apparatus of any one of claims 7 to 11, wherein the apparatus further comprises:
the third acquisition module is used for responding to the starting operation of a user and pulling the stored training samples;
The model updating module is used for updating the model parameters of the user interest degree prediction model by utilizing the stored training samples;
the first prediction module includes:
and the first prediction sub-module is used for obtaining the user interestingness corresponding to the plurality of search results through the updated user interestingness prediction model.
13. An electronic device, comprising: a processor, a memory and a computer program stored on the memory and executable on the processor, which when executed by the processor performs the steps of the search result presentation method as claimed in any one of claims 1 to 6.
14. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the search result presentation method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211733384.9A CN116108269A (en) | 2022-12-30 | 2022-12-30 | Search result display method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211733384.9A CN116108269A (en) | 2022-12-30 | 2022-12-30 | Search result display method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116108269A true CN116108269A (en) | 2023-05-12 |
Family
ID=86257481
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211733384.9A Pending CN116108269A (en) | 2022-12-30 | 2022-12-30 | Search result display method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116108269A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116595256A (en) * | 2023-05-22 | 2023-08-15 | 毕加展览有限公司 | Method and system for data screening and immersive interaction of digital exhibition |
-
2022
- 2022-12-30 CN CN202211733384.9A patent/CN116108269A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116595256A (en) * | 2023-05-22 | 2023-08-15 | 毕加展览有限公司 | Method and system for data screening and immersive interaction of digital exhibition |
| CN116595256B (en) * | 2023-05-22 | 2023-11-03 | 毕加展览有限公司 | Method and system for data screening and immersive interaction of digital exhibition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11765246B2 (en) | Topical activity monitor and identity collector system | |
| US7941383B2 (en) | Maintaining state transition data for a plurality of users, modeling, detecting, and predicting user states and behavior | |
| US9069872B2 (en) | Relating web page change with revisitation patterns | |
| US9171078B2 (en) | Automatic recommendation of vertical search engines | |
| US7818208B1 (en) | Accurately estimating advertisement performance | |
| JP5572596B2 (en) | Personalize the ordering of place content in search results | |
| US8374985B1 (en) | Presenting a diversity of recommendations | |
| US8332775B2 (en) | Adaptive user feedback window | |
| US20090089293A1 (en) | Selfish data browsing | |
| US9529915B2 (en) | Search results based on user and result profiles | |
| US20140280890A1 (en) | Method and system for measuring user engagement using scroll dwell time | |
| US20070143300A1 (en) | System and method for monitoring evolution over time of temporal content | |
| EP2407897A1 (en) | Device for determining internet activity | |
| CN109168047B (en) | Video recommendation method and device, server and storage medium | |
| US20200286154A1 (en) | Utilizing item-level importance sampling models for digital content selection policies | |
| WO2016187705A1 (en) | System and method for ranking search results | |
| US9594809B2 (en) | System and method for compiling search results using information regarding length of time users spend interacting with individual search results | |
| EP2541445A2 (en) | A method and a system for analysing impact of changes to content of a website | |
| CN111400586A (en) | Group display method, terminal, server, system and storage medium | |
| KR101659064B1 (en) | Method and apparatus for calculating contents evaluation scores by using user feedbacks | |
| TWI417751B (en) | Information providing device, information providing method, information application program, and information recording medium | |
| US10296924B2 (en) | Document performance indicators based on referral context | |
| CN116108269A (en) | Search result display method and device, electronic equipment and storage medium | |
| TWI399657B (en) | A provider, a method of providing information, a program, and an information recording medium | |
| JP5813052B2 (en) | Information processing apparatus, method, and program |
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 |