CN110648157B - Method, device, terminal and storage medium for determining reserve price in advertisement market - Google Patents
Method, device, terminal and storage medium for determining reserve price in advertisement market Download PDFInfo
- Publication number
- CN110648157B CN110648157B CN201810682467.7A CN201810682467A CN110648157B CN 110648157 B CN110648157 B CN 110648157B CN 201810682467 A CN201810682467 A CN 201810682467A CN 110648157 B CN110648157 B CN 110648157B
- Authority
- CN
- China
- Prior art keywords
- reserve price
- keyword
- mining
- determining
- similarity
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0283—Price estimation or determination
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0249—Advertisements based upon budgets or funds
Abstract
The embodiment of the invention provides a method, a device, a terminal and a storage medium for determining a reserve price in an advertisement market, wherein the method comprises the following steps: acquiring a keyword set and a mining word set in the historical advertisement, wherein the keyword set comprises at least one keyword in the historical advertisement, and the mining word set is formed by the keywords mined according to the click rate in the historical advertisement; determining an optimal reserve price corresponding to the keywords in the keyword set and a reserve price of the mining words corresponding to the keywords in the mining word set; and determining a final reserve price corresponding to the keyword according to the optimal reserve price and the mining word reserve price. According to the technical scheme provided by the invention, the final reserve price is determined through the optimal reserve price and the reserve price of the mining word, and the final reserve price comprises the reserve price of the mining word, so that the range of the keywords covered by the final reserve price is effectively expanded, the quality effect of the advertisement seen by a user is ensured, and the benefit of a search engine platform is also ensured.
Description
Technical Field
The present invention relates to the field of data processing technologies, and in particular, to a method, an apparatus, a terminal, and a storage medium for determining a reserve price in an advertisement market.
Background
The search bidding advertisement is an important type in an online advertisement market, the search bidding advertisement is audience-oriented by taking a current query word (query) searched in real time as granularity, and is sold and presented for thousands of times for charging according to a bidding mode, a bid object for bidding the search advertisement is a bidding keyword, for an advertiser, the advertiser can purchase a plurality of bidding keywords (bid), different keyword advertisers can bid with different prices, for a search engine advertisement platform, a search engine matches a plurality of bidding keywords (bid) through a search word (query) input by a user, each keyword has a plurality of advertiser bidding purchases, and the search engine comprehensively sorts the keywords according to matching precision, advertiser bids, advertiser advertisement and search word relevance, and selects a plurality of advertisements suitable for presentation to be presented to the user. The broad second-order bidding strategy is mostly adopted in the search advertisement engine industry, that is, when a plurality of advertisement positions are searched at one time, each advertiser winning a position is charged according to the bid of the advertiser of the next position.
However, the traffic distribution of some bidding keywords is in the long-tailed part of the search traffic, the bidding is not sufficient, the commercial value of the search query itself is high, the bidding price and the quality degree are uneven, the effect of presenting the advertisement is poor if the price threshold is not set, or the revenue loss of the engine platform is caused when the advertiser bids the bidding keywords at a very low price. The bidding advertising marketplace typically sets a minimum bid to win the auction position, the market reserve price, which is the minimum bid that the advertiser can bid upon entering the final rank queue in the search engine mechanism implementation.
At present, the method for setting reserve price in the industry usually calculates reserve price for query with stable bidding queue by greedy algorithm according to the industry, price and severity of the bidding queue, but the method has the following disadvantages: 1) Only calculating the reserve price of the query with the stable bidding queue covers part of the flow, and neglecting the high commercial value query distributed in the long tail flow; 2) The price computed from the advertiser's bid and greedy algorithm is often only valid for the current search and current ad queue, and the advertiser's bid is not constant, so the traditional industry computed reserve price is not generalised.
Disclosure of Invention
The embodiment of the invention provides a method, a device, a terminal and a storage medium for determining a reserve price in an advertisement market, which are used for solving the problems or other potential problems in the prior art.
The first aspect of the embodiments of the present invention provides a method for determining a reserve price in an advertisement market, including:
acquiring a keyword set and a mining word set in a historical advertisement, wherein the keyword set comprises at least one keyword in the historical advertisement, and the mining word set is formed by the keywords mined according to click rate in the historical advertisement;
determining an optimal reserve price corresponding to the keywords in the keyword set and a reserve price of the mining words corresponding to the keywords in the mining word set;
and determining a final reserve price corresponding to the keyword according to the optimal reserve price and the reserve price of the mining word.
The method for determining the reserve price of the mining words corresponding to the keywords in the mining word set comprises the following steps:
and determining the reserve price of the mining words corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set.
The method for determining the reserve price of the mining word corresponding to the mining word set according to the keyword set, the optimal reserve price corresponding to the keywords in the keyword set and the mining word set comprises the following steps:
acquiring the similarity between the keyword set and the mining word set;
and determining the reserve price of the mining words according to the similarity and the optimal reserve price.
The method for acquiring the similarity between the keyword set and the mining word set comprises the following steps:
vectorizing the keyword set and the mining word set by using a hidden Dirichlet distribution model to obtain the vectorized keyword set and mining word set;
acquiring vectorization similarity of the keyword set subjected to vectorization processing and the mining word set;
and determining the vectorization similarity as the similarity between the keyword set and the mining word set.
The method for determining the reserve price of the mining word according to the similarity and the optimal reserve price comprises the following steps:
multiplying the similarity between at least one keyword in the keyword set and the mining word set by the optimal reserve price corresponding to the keyword to obtain at least one intermediate reserve price;
accumulating the similarity between at least one keyword in the keyword set and the mining word set to obtain a similarity sum value;
and determining the ratio of the sum of all the intermediate reserve prices to the similarity sum value as the reserve price of the mining word.
A second aspect of the embodiments of the present invention provides an apparatus for determining a reserve price in an advertisement market, including:
the system comprises an acquisition module, a search module and a search module, wherein the acquisition module is used for acquiring a keyword set and a mining word set in the historical advertisement, the keyword set comprises at least one keyword in the historical advertisement, and the mining word set is formed by the keywords mined in the historical advertisement according to click rate;
the determining module is used for determining the optimal reserve price corresponding to the keywords in the keyword set and the reserve price of the mining words corresponding to the keywords in the mining word set;
and the processing module is used for determining the final reserve price corresponding to the keyword according to the optimal reserve price and the reserve price of the mining word.
The apparatus as described above, the determining means to:
and determining the reserve price of the mining words corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set.
The apparatus as described above, the determining means to:
acquiring the similarity between the keyword set and the mining word set;
and determining the reserve price of the mining words according to the similarity and the optimal reserve price.
The apparatus as described above, the determining means to:
vectorizing the keyword set and the mining word set by using a hidden Dirichlet distribution model to obtain a vectorized keyword set and a vectorized mining word set;
acquiring vectorization similarity of the keyword set subjected to vectorization processing and the mining word set;
and determining the vectorization similarity as the similarity between the keyword set and the mining word set.
The apparatus as described above, the determining means to:
multiplying the similarity between at least one keyword in the keyword set and the mining word set by the optimal reserve price corresponding to the keyword to obtain at least one intermediate reserve price;
accumulating the similarity between at least one keyword in the keyword set and the mining word set to obtain a similarity sum value;
and determining the ratio of the sum of all the intermediate reserve prices to the similarity sum value as the reserve price of the mining word.
A third aspect of an embodiment of the present invention provides a terminal for determining a reserve price in an advertisement market, including:
a memory;
a processor; and
a computer program;
wherein the computer program is stored in the memory and configured to be executed by the processor to implement a method of reserve price determination in an advertising market as described in the first aspect.
A fourth aspect of the embodiments of the present invention provides a storage medium, which is a computer-readable storage medium having a computer program stored thereon;
the computer program is executed by a processor to implement a method of determining a reserve price in an advertising market as described in the first aspect.
According to the method, the device, the terminal and the storage medium for determining the reserve price in the advertisement market, provided by the embodiment of the invention, the optimal reserve price and the reserve price of the mining word are determined, then the final reserve price can be determined according to the optimal reserve price and the reserve price of the mining word, and after the final reserve price is obtained, the final reserve price can be applied to an online reserve price dictionary; because the final reserve price given by the embodiment includes the mining word reserve price, the scope of the keywords covered by the final reserve price is effectively expanded, so that the quality effect of the advertisement seen by the user is ensured, the income of a search engine platform is also ensured, the practicability of the method is further improved, and the popularization and the application of the market are facilitated.
Drawings
FIG. 1 is a flow chart illustrating a method for determining a reserve price in an advertising marketplace according to an embodiment of the present invention;
fig. 2 is a schematic flow chart illustrating a process of determining a retention price of a mining word corresponding to the mining word set according to the keyword set, the optimal retention price, and the mining word set according to the embodiment of the present invention;
fig. 3 is a schematic flowchart of a process of obtaining similarity between the keyword set and the mining word set according to the embodiment of the present invention;
fig. 4 is a schematic flowchart of determining a reserve price of the mining word according to the similarity and the optimal reserve price according to the embodiment of the present invention;
FIG. 5 is a flowchart illustrating a method for determining a reserve price in an advertisement market according to an embodiment of the present invention;
FIG. 6 is a schematic structural diagram of an apparatus for determining a reserve price in an advertisement market according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a terminal for determining a reserve price in an advertisement market according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. 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.
The terms "comprises" and "comprising," and any variations thereof, in the description and claims of this invention, are intended to cover non-exclusive inclusions, e.g., a process or an apparatus that comprises a list of steps is not necessarily limited to those structures or steps expressly listed but may include other steps or structures not expressly listed or inherent to such process or apparatus.
Fig. 1 is a schematic flow chart of a method for determining a reserve price in an advertisement market according to an embodiment of the present invention, and referring to fig. 1, the embodiment provides a method for determining a reserve price in an advertisement market, the method being used for accurately and broadly predicting and evaluating a reserve price in an advertisement market, and in particular, the method includes:
s1: acquiring a keyword set and a mining word set in a historical advertisement, wherein the keyword set comprises at least one keyword in the historical advertisement, and the mining word set is formed by the keywords mined according to click rate in the historical advertisement;
the main source of the keywords in the keyword set is the result of analyzing and processing the display logs of the historical advertisements; the mining word set may be determined according to the click rate, for example, a keyword with a higher click rate (the click rate is greater than or equal to a preset click rate threshold value) is obtained, at this time, the keyword may be used as a component element in the mining word set, it should be noted that the component element in the mining word set may be present in the keyword set, or the component element may not be present in the keyword set, which is preferable.
S2: determining an optimal reserve price corresponding to the keywords in the keyword set and a reserve price of the mining words corresponding to the keywords in the mining word set;
the mining word reserve price corresponding to the mining word set corresponds to a word in the mining word set, and one mining word set can correspond to one or more mining word reserve prices; after the keyword set and the mining word set are obtained, the optimal reserve price corresponding to the keywords in the keyword set can be determined by using the mapping relation between the keyword set and the optimal reserve price, and the mining word reserve price corresponding to the keywords in the mining word set can be determined by using the mapping relation between the mining word set and the mining word reserve price.
S3: and determining a final reserve price corresponding to the keyword according to the optimal reserve price and the reserve price of the mining word.
After the optimal reserve price and the reserve price of the mining word are obtained, the final reserve price corresponding to the keyword may be determined according to a preset processing strategy, for example, the final reserve price of the keyword may be set as an average value, a sum value, a difference value, or the like of the optimal reserve price of the keyword and the reserve price of the mining word. Of course, those skilled in the art may also set the reservation price according to specific design requirements, as long as the final reservation price can be determined by synthesizing the optimal reservation price and the reserve price of the mining word, which is not described herein again.
In the method for determining the reserve price in the advertisement market provided by the embodiment, the optimal reserve price and the reserve price of the mining word are determined, then the final reserve price can be determined according to the optimal reserve price and the reserve price of the mining word, and after the final reserve price is obtained, the final reserve price can be applied to an online reserve price dictionary; because the final reserve price given by the embodiment includes the mining word reserve price, the scope of the keywords covered by the final reserve price is effectively expanded, so that the quality effect of the advertisement seen by the user is ensured, the income of a search engine platform is also ensured, the practicability of the method is further improved, and the popularization and the application of the market are facilitated.
Fig. 2 is a schematic flow chart illustrating a process of determining a retention price of a mining word corresponding to the mining word set according to the keyword set, the optimal retention price, and the mining word set according to the embodiment of the present invention; fig. 3 is a schematic flowchart of a process of obtaining similarity between the keyword set and the mining word set according to the embodiment of the present invention; fig. 4 is a schematic flowchart of determining a reserve price of the mining word according to the similarity and the optimal reserve price according to the embodiment of the present invention; on the basis of the foregoing embodiment, with reference to fig. 1 to 4, it can be seen that, in this embodiment, a specific implementation process for determining a reserve price of a mining word corresponding to a keyword in a mining word set is not limited, and a person skilled in the art may set the reserve price according to a specific design requirement, and preferably, the determining a reserve price of a mining word corresponding to a keyword in a mining word set in this embodiment may include:
s21: and determining the reserve price of the mining words corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set.
Specifically, determining the reserve price of the mining word corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set may include:
s211: acquiring the similarity between the keyword set and the mining word set;
obtaining the similarity between the keyword set and the mining word set may include:
s2111: vectorizing the keyword set and the mining word set by using a hidden Dirichlet distribution model to obtain the vectorized keyword set and mining word set;
after the keyword set and the mining word set are obtained, a language model based on the advertisement corpus needs to be constructed so as to realize theme feature expression on the keyword set and the mining word set. After the features and trial scenes of the topic models (such as word2Vec, RNN, param 2Vec, and the like) are investigated and compared, a latent dirichlet distribution (LDA) model is selected as a semantic vectorization model in the embodiment, on one hand, because the keyword set and the mining word set are formed by splicing description, title, and bid keywords, and a part of sequential semantic features are lost, the semantic model requiring the semantic sequential features may no longer be suitable. On the other hand, LDA has mature application in both the industry and the academia, and the importance distribution of each word under each topic category can be obtained, so that the model can be better understood and optimized.
S2112: acquiring vectorization similarity of the keyword set subjected to vectorization processing and the mining word set;
specifically, the cos cosine can be used to calculate the similarity between the vectorized keyword set and the mining word set.
S2113: and determining the vectorization similarity as the similarity between the keyword set and the mining word set.
S212: and determining the reserve price of the mining words according to the similarity and the optimal reserve price.
And determining the reserve price of the mining word according to the similarity and the optimal reserve price may include:
s2121: multiplying the similarity between at least one keyword in the keyword set and the mining word set by the optimal reserve price corresponding to the keyword to obtain at least one intermediate reserve price;
s2122: accumulating the similarity between at least one keyword in the keyword set and the mining word set to obtain a similarity sum value;
s2123: and determining the ratio of the sum of all the intermediate reserve prices to the similarity sum value as the reserve price of the mining word.
For example: the keyword set comprises a first keyword, a second keyword and a third keyword, the similarity between the keyword and a certain word in the mining word set is respectively similarity 1, similarity 2 and similarity 3, and then the reserved price of the mining word corresponding to the word in the mining word set is: (first keyword x similarity 1+ second keyword x similarity 2+ third keyword x similarity 3)/similarity 1+ similarity 2+ similarity 3. And the similarity 1 of the first keyword, the similarity 2 of the second keyword and the similarity 3 of the third keyword are all intermediate reserve prices.
By acquiring the reserve price of the mining word in the mode, the accuracy and the reliability of determining the reserve price of the mining word are effectively ensured, and the use reliability of the method is further improved.
Further, in order to improve the practicability of the method, the embodiment describes a specific implementation manner of the optimal reserve price, which may specifically be implemented by the following steps:
s101: obtaining bidding queue information of keywords in historical advertisements;
the historical advertisement can comprise a plurality of keywords, each keyword has corresponding bidding queue information, when the bidding queue information is obtained, the corresponding keyword can be determined according to the historical advertisement, specifically, the keyword can be determined according to log information in the historical advertisement, the keyword can be at least one word in a query statement, and certainly, the keyword can also be the whole query statement; after determining the keywords in the historical advertisement, the bidding queue information corresponding to the keywords may be obtained using a preset mapping relationship.
S102: analyzing and processing the bidding queue information according to a preset Gaussian mixture model to obtain price distribution information;
after the bid queue information is obtained, analysis processing needs to be performed on the bid queue information to obtain price distribution information, specifically, the bid queue information may be processed by using a gaussian mixture model, where the gaussian mixture model is arbitrarily set by a user according to requirements, for example, the gaussian mixture model may be two gaussian mixture models or three gaussian mixture models, and the gaussian mixture model may perform tuning processing on the bid queue information, so that price distribution information corresponding to a keyword may be obtained.
S103: and analyzing and processing the price distribution information according to a balanced value theory to obtain the optimal reserved price corresponding to the keyword in the advertisement market.
After the price distribution information is obtained, the price distribution information can be analyzed and processed by using a balanced value theory, so that an optimal reserve price corresponding to the keyword in the advertisement market can be obtained, wherein the optimal reserve price is a threshold price bid by an advertiser and a charging price of a generalized second-order bid exhibited by the last advertisement.
The method for determining the reserve price in the advertisement market provided by the embodiment obtains the bid queue information, analyzes and processes the bid queue information by using the Gaussian mixture model to obtain the price distribution information, further analyzes and processes the price distribution information according to the equilibrium value theory to obtain the optimal reserve price corresponding to the keyword in the advertisement market, effectively overcomes the defect that the reserve price in the prior art is not wide, and not only ensures the quality effect of the advertisement seen by a user, but also ensures the benefit of a search engine platform through the determination of the optimal reserve price, thereby improving the practicability of the method and being beneficial to the popularization and application of the market.
Further, obtaining bid queue information for keywords in historical advertisements may include:
s1011: acquiring a display log of a window of a historical advertisement in a preset time period;
the preset time period is set by a user at will according to requirements, and can be one month, a half month, or 5 days, 10 days and the like; the bidding queue information can be accurately determined by obtaining the display log of the window within the preset time period.
S1012: and determining bidding queue information of the keywords according to the display log.
After the display log is obtained, the bid queue information of the keyword may be determined according to the display log, and specifically, determining the bid queue information of the keyword according to the display log may include:
s10121: determining original bidding queue information of the keywords according to the display log;
by utilizing the display logs of the windows with a period of historical accumulation, the display advertisement bids of the keywords can be mined, and the keywords are used as keys to aggregate the historical advertisement bidding queues of the keywords and used as the original bidding queue information of the fitting distribution.
S10122: and carrying out singular value filtering and frequency screening processing on the original bidding queue information to obtain the bidding queue information of the keyword.
After the original bidding queue information is obtained, in order to ensure the accuracy and reliability of the obtaining of the bidding queue information, singular value filtering and frequency screening need to be performed on the original bidding queue information according to empirical values, so as to obtain data which is really input to the training of the Gaussian mixture model, namely, the bidding queue information of the keyword.
The bidding queue information of the keywords is acquired through the process, so that the accuracy and the reliability of acquiring the bidding queue information are effectively ensured, and the accuracy of the method is further improved.
Further, analyzing the bidding queue information according to a preset gaussian mixture model to obtain price distribution information may include:
s1021: carrying out logarithmic transformation processing on the bidding queue information to obtain target bidding queue information;
in the process of fitting the mixed gaussian distribution, the most appropriate mixed gaussian number needs to be found, and if the bidding queue information (advertiser bids) is directly used as a training data point to fit the distribution, the fluctuation range is too large, so that the value range is out of the computer processing numerical range, and therefore, in order to avoid the situation, log transformation processing is performed on each sample point in the bidding queue information in the embodiment, so that the target bidding queue information can be obtained.
S1022: and analyzing and processing the target bidding queue information according to the Gaussian mixture model to obtain price distribution information.
After obtaining the information of the target bidding queue, analyzing the information of the target bidding queue by using a gaussian mixture model to obtain price distribution information, specifically, by observing data distribution and combining historical experience accumulation of the information of the target bidding queue, in this embodiment, two gaussian mixture distributions are selected to fit the bidding distribution queue, and then, analyzing the information of the target bidding queue according to the gaussian mixture model to obtain the price distribution information may include:
s10221: analyzing and processing the target bidding queue information by using the following formula to obtain price distribution information:
wherein alpha is 1 Is a first preset fitting weight of two preset Gaussian distributions, alpha 2 Is a second preset fitting weight of two preset Gaussian distributions, mu 1 Is a first preset mean value mu which is obtained by fitting two preset Gaussian distributions according to target bidding queue information 2 Is a second preset mean value, sigma, which is obtained by fitting two preset Gaussian distributions according to the target bidding queue information 1 ) Is a first preset fluctuation parameter, sigma, which is obtained by fitting two preset Gaussian distributions according to target bidding queue information 2 ) And the second preset fluctuation parameter is a second preset fluctuation parameter which is obtained by fitting two preset Gaussian distributions according to the target bidding queue information, wherein S is the target bidding queue information, and f (S) is the price distribution information.
The price distribution information is determined through the two Gaussian mixture distributions, so that the accuracy and the reliability of the determination of the price distribution information are effectively ensured, and the practicability of the method is further improved.
Further, on the basis of the foregoing embodiment, this embodiment does not limit an implementation manner of determining the optimal reserve price, and a person skilled in the art may set the optimal reserve price according to a specific design requirement, and preferably, the analyzing and processing the price distribution information according to the equilibrium value theory in this embodiment to obtain the optimal reserve price of the keyword in the advertisement market may include:
s1031: acquiring a cumulative distribution function corresponding to the price distribution information;
s1032: analyzing and processing the price distribution information according to a balance value theory and an accumulated distribution function by using the following formula to obtain the optimal reserve price:
wherein, the first and the second end of the pipe are connected with each other,s is the optimum reserve price, f (S) * ) For price distribution information, F (S) * ) Is a cumulative distribution function of the price distribution information.
The optimal reserve price is determined through the equilibrium value theory, the accuracy and the reliability of determining the optimal reserve price are effectively guaranteed, and the practicability of the method is further improved.
In specific application, the method provided by the embodiment of the application establishes semantic topic relevance relation between massive search queries (representing query sentences recorded by a search log) and fiercely bidded base queries (base query sets), so that more queries with high commercial value are mined, similarity between the predict query sets and the base query sets is estimated according to calculated topic vectors of a language model, and then semantically relevant reserve prices are estimated. On the other hand, the concept of the distribution of the bids of the advertisers is provided according to the Nash equilibrium bidding auction theory and the bidding mechanism design under the Myrson equilibrium condition, and the problem of calculating the reserve price is converted into the problem of dynamic planning of the reserve price with known keyword bid distribution and maximum income. And establishing continuous probability distribution for the bidding of the keywords through a Gaussian mixture model, and calculating a reserve price through estimating the bidding distribution of the keywords.
Referring to fig. 5, the method provided by the embodiment of the present application is divided into three blocks as a whole: GMM price distribution prediction, LDA model word expansion and online query reserve price application.
The first part of GMM price distribution prediction is that aggregation is carried out on a bid queue of keywords, a search engine data log of a time window is used, parameters of mixed Gaussian distribution are adjusted based on dynamic sample data, meanwhile, strategies such as frequency threshold value and singular value filtering are set, the problem of overfitting is solved, a bid distribution function of the keywords is finally obtained, the optimal reserve price is calculated and applied to a real-time advertisement bidding strategy by combining the Myrson game theory on line, and finally the income of the obtained eCPM is increased by more than 2 percent.
Specifically, the GMM price distribution prediction part is integrally divided into three blocks, namely log accumulation and data mining, training and tuning of a Gaussian mixture distribution model, and calculation and application of the on-line optimal reserve price. And fitting the historical accumulated bidding queue of the bidding keywords with sufficient bidding by using mixed Gaussian distribution, and calculating the optimal reserve price offline according to the Myrson balanced bidding theory so as to achieve the win-win situation of the user and the engine platform. The method comprises the following specific steps:
step 1: log accumulation and data mining
In order to fit accurate price distribution, sufficient bidding queue information is needed, a display log of a time window is accumulated by using history, display advertisement bids of keywords are mined, the keywords are used as historical advertisement bidding queues of the keywords which are aggregated by using the keys as keys and used as original data (namely original bidding queue information) of the fitting distribution, and singular value filtering and frequency screening are carried out on the original queues according to empirical values. Data that is really input (bid queue information) to the gaussian mixture model training is obtained.
Step 2: model training and tuning
The embodiment selects two gaussian mixture distributions to fit the bid distribution queue, which can be specifically referred to the implementation manner of steps S1021-S1022 and formula (1).
And step 3: reserve price calculation and application
After the price distribution information is obtained, the probability distribution function is integrated to obtain the cumulative distribution function of the price distribution information, and then the formula (2) is solved to obtain the optimal reserve price of the keyword.
The method is online and applied to an advertisement bidding frame of a real-time search engine in a word list mode, when the retrieved query is matched with the keyword optimal reserve price word list, the optimal reserve price s is a threshold price bid by an advertiser and a charging price of generalized second-order bidding presented by the last advertisement, and therefore the quality effect of the advertisement seen by a user is guaranteed, and the benefit of a search engine platform is also guaranteed.
In the second part, the word expansion of the LDA topic model mainly involves: storing and processing data and linguistic data of a base query set (namely the keyword set) and a prediction query (namely the mining word set), training an LDA model according to the linguistic data and the data of the first part, and finally, optimizing and evaluating the model to give theme vectors of the base query and the prediction query; and expanding a new word list according to the similarity.
Specifically, the LDA word expansion part is integrally divided into two steps:
step 1: base query set and predict query data storage and corpus processing
The base query mainly comes from the query obtained by the distribution of the first part of GMMs, the optimal reserve price corresponding to the keywords in the base query is mainly determined by the first part of GMMs, the predict query is a new query set which needs to be expanded, the base query has high commercial value, but a batch of queries with high click rate are mined as the predict query set according to historical click data because the bid queue is insufficient or the historical search flow is unstable and the batch of queries with high click rate cannot enter the base query set.
On the basis of obtaining a base query set and a prediction query set, corpus data of the LDA needs to be trained, based on a corresponding search engine, because the base query and the prediction query are both queries with higher click rate, descriptions and titles of advertisements displayed by the queries are mined according to click logs, then aggregation is performed according to the queries, each query forms a corpus document needed by the LDA, and the corpus document is used as corpus storage of next-step model training after word segmentation.
Step 2: model tuning and evaluation
On the basis of the linguistic data in the step 1, a language model based on the advertisement linguistic data needs to be constructed, and theme feature expression is carried out on the query. Specifically, a hidden dirichlet distribution (LDA) model is used as a semantic vectorization model, so that the hidden dirichlet distribution model can be used for vectorizing the keyword set and the mining word set to obtain the vectorized keyword set and mining word set.
It should be noted that, in this embodiment, the LDA topic model may also be optimized by using a Gibbs sampling method based on a commercialized corpus of a search engine, where the LDA topic model requires three parameters, namely, a document-topic hidden parameter alpha, a topic-word hidden parameter beta, and a topic number k, and according to an empirical value and an advertisement corpus characteristic, the more dispersed topics are that the alpha should not be too small and the beta should be set smaller, and results obtained when the alpha =0.5 and the beta =0.1 after multiple times of experimental parameter tuning are already in line with expectations. Whereas the topoc number k takes 300, in principle the larger the better. Thus, the topic vectors of the prediction query and the base query are obtained, and the vectorization of the query is completed.
And the third part, a base query reserve price obtained by a new reserve price word list and price distribution gives a reasonable reserve price of the query, and the reserve price is applied to an online query reserve price dictionary.
Specifically, the obtained reserve price of the query is to calculate the similarity between the base query and the prediction query to expand a new reserve price vocabulary, on the basis of the topic vector obtained in the LDA word expansion step 2, the cos cosine is used to calculate the similarity between the prediction query and each base query set, the top N (N may be 2, 3, 4, etc.) cos cosine with the largest base query and price are reserved for each prediction query, and the price of the prediction query is smoothed, so that the dictionary expansion and price prediction are realized.
On the other hand, the LDA word expansion part greatly improves the flow coverage of the reserve price word list, and is a new exploration direction of the topic model on the search advertisement service. Taking graph 1 as an example, the "airline ticket in krama to zheng" is the query in the predict query set, and "which website is the cheapest to buy the airline ticket", "premium airline ticket booking official website", "hotel booking" is the query in the base query set, and specifically, after obtaining semantic vectors of the above four queries by LDA, calculating cos distance to obtain similarity, for example, the similarity of predict query and three base queries is 0.9, 0.8, 0.7, and the reserve price of predict query is equal to (0.9 + reserve price 1+0.8 + reserve price 2+0.7 + reserve price 3)/(0.9 +0.8+ 0.7).
According to the similarity ranking of the theme vectors, the query with the highest similarity to the 'airline tickets of the clara to zheng state' is the three queries, and the query reserve price with high similarity of the theme vectors can be intuitively seen to be very close.
TABLE 1
| Air ticket of clamayi to zheng | |
| It is cheapest to buy the air ticket to which website | 231.13 |
| Special-price air ticket booking official website | 241.132 |
| Hotel reservation | 254.069 |
In an online small flow experiment, a base query set can obtain 6 percent increase of eCPM, and the expanded prediction query set also obtains 6 percent income increase, which shows that language vectors have good effect on price prediction and dictionary expansion, and also give good interpretability on the aspect of semantics.
The implementation effect of the method can be evaluated according to the income of an engine platform and the click condition effect of a user on the advertisement, and is specific, when a price distribution model predicted by GMM is tested on line, the income of eCPM 2 points is increased, the income of the platform is increased by the price strategy, the click rate of the user on the advertisement is increased by 1 point, the price strategy is proved to also improve the user experience, the practicability of the method is further ensured, and the method is favorable for popularization and application of the market.
FIG. 6 is a schematic structural diagram of an apparatus for determining a reserve price in an advertisement market according to an embodiment of the present invention; as can be seen with reference to fig. 6, the present embodiment provides an apparatus for determining a reserve price in an advertisement market, the apparatus being configured to perform the above-mentioned determining method, and in particular, the apparatus may include:
the system comprises an acquisition module 1, a search module and a search module, wherein the acquisition module is used for acquiring a keyword set and a mining word set in a historical advertisement, the keyword set comprises at least one keyword in the historical advertisement, and the mining word set is composed of keywords mined according to click rate in the historical advertisement;
the determining module 2 is used for determining the optimal reserve price corresponding to the keywords in the keyword set and the reserve price of the mining words corresponding to the keywords in the mining word set;
and the processing module 3 is used for determining a final reserve price corresponding to the keyword according to the optimal reserve price and the reserve price of the mining word.
In this embodiment, specific shape structures of the obtaining module 1, the determining module 2, and the processing module 3 are not limited, and those skilled in the art may arbitrarily set the obtaining module, the determining module, and the processing module according to the implemented function, and details are not described herein again; in addition, in this embodiment, the specific implementation process and implementation effect of the operation steps implemented by the obtaining module 1, the determining module 2, and the processing module 3 are the same as the specific implementation process and implementation effect of the steps S1 to S3 in the foregoing embodiment, and the above statements may be specifically referred to, and are not repeated herein.
When the determining module 2 determines the reserve price of the mining word corresponding to the keyword in the mining word set, the determining module 2 is specifically configured to: and determining the reserve price of the mining words corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set.
Further, when the determining module 2 determines the reserve price of the mining word corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set, the determining module 2 is specifically configured to perform: acquiring the similarity between the keyword set and the mining word set; and determining the reserve price of the mining words according to the similarity and the optimal reserve price.
When the determining module 2 obtains the similarity between the keyword set and the mining word set, the determining module 2 is specifically configured to: vectorizing the keyword set and the mining word set by using a hidden Dirichlet distribution model to obtain a vectorized keyword set and a vectorized mining word set; acquiring vectorization similarity of the vectorization-processed keyword set and the mining word set; and determining the vectorization similarity as the similarity between the keyword set and the mining word set.
Further, when the determining module 2 determines the reserve price of the mining word according to the similarity and the optimal reserve price, the determining module 2 is specifically configured to perform: multiplying the similarity between at least one keyword in the keyword set and the mining word set by the optimal reserve price corresponding to the keyword to obtain at least one intermediate reserve price; accumulating the similarity between at least one keyword in the keyword set and the mining word set to obtain a similarity sum value; and determining the ratio of the sum of all the intermediate reserve prices to the similarity sum value as the reserve price of the mining word.
The device for determining a reserve price in an advertisement market provided by this embodiment can be used to execute the method corresponding to the embodiments in fig. 2 to 5, and the specific execution manner and the beneficial effects thereof are similar, and are not described herein again.
Another aspect of the present embodiment provides a terminal for determining a reserve price in an advertisement market, including:
a memory;
a processor; and
a computer program;
wherein the computer program is stored in the memory and configured to be executed by the processor to implement a method of determining reserve prices in an advertising marketplace as described above.
Specifically, fig. 7 is a schematic structural diagram of a terminal for determining a reserve price in an advertisement market according to an embodiment of the present invention.
As shown in fig. 7, the determination terminal 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814, and a communication component 816.
The processing component 802 generally controls the overall operation of the determination terminal 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing components 802 may include one or more processors 820 to execute instructions to perform all or a portion of the steps of the methods described above. Further, the processing component 802 can include one or more modules that facilitate interaction between the processing component 802 and other components. For example, the processing component 802 can include a multimedia module to facilitate interaction between the multimedia component 808 and the processing component 802.
The memory 804 is configured to store various types of data to support operations at the determination terminal 800. Examples of such data include instructions for any application or method operating on the determination terminal 800, contact data, phonebook data, messages, pictures, videos, and the like. The memory 804 may be implemented by any type or combination of volatile or non-volatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.
The multimedia component 808 includes a screen providing an output interface between the determination terminal 800 and the user. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.
The audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a Microphone (MIC) configured to receive an external audio signal when the terminal 800 is determined to be in an operating mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may further be stored in the memory 804 or transmitted via the communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.
The I/O interface 812 provides an interface between the processing component 802 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to: a home button, a volume button, a start button, and a lock button.
The communication component 816 is configured to facilitate determining a wired or wireless manner of communication between the terminal 800 and other devices. It is determined that the terminal 800 can access a wireless network based on a communication standard, such as WiFi,2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, communications component 816 further includes a Near Field Communications (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.
In an exemplary embodiment, the determination terminal 800 may be implemented by one or more Application Specific Integrated Circuits (ASICs), digital Signal Processors (DSPs), digital Signal Processing Devices (DSPDs), programmable Logic Devices (PLDs), field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components for performing the above-described methods.
Another aspect of the embodiments of the present invention provides a storage medium, which is a computer-readable storage medium having a computer program stored thereon; a computer program is executed by a processor to implement a method of reserve price determination in an advertising marketplace as described above.
Finally, it should be noted that, as one of ordinary skill in the art will appreciate, all or part of the processes of the methods of the embodiments described above may be implemented by hardware related to instructions of a computer program, where the computer program may be stored in a computer-readable storage medium, and when executed, the computer program may include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), a Random Access Memory (RAM), or the like.
Each functional unit in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a separate product, may also be stored in a computer readable storage medium. The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
The above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (12)
1. A method for determining a reserve price in an advertising marketplace, the method performed by a processor, comprising:
acquiring a keyword set and a mining word set in a historical advertisement, wherein the keyword set comprises at least one keyword in the historical advertisement, the keyword in the keyword set is derived from a result of analyzing and processing a display log of the historical advertisement, and the mining word set is formed by the keywords mined according to a click rate in the historical advertisement;
determining an optimal reserve price corresponding to the keywords in the keyword set and a reserve price of the mining words corresponding to the keywords in the mining word set;
determining a final reserve price corresponding to the keyword according to the optimal reserve price and the reserve price of the mining word;
applying the final reserve price to a reserve price dictionary on the line.
2. The method of claim 1, wherein determining a mining word reserve price corresponding to a keyword in the set of mining words comprises:
and determining the reserve price of the mining words corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set.
3. The method of claim 2, wherein determining the reserve price of the mining word corresponding to the set of mining words according to the set of keywords, the optimal reserve price and the set of mining words comprises:
acquiring similarity between words in the keyword set and words in the mining word set;
and determining the reserve price of the mining words according to the similarity and the optimal reserve price.
4. The method of claim 3, wherein obtaining similarity between words in the set of keywords and words in the set of mining words comprises:
vectorizing the keyword set and the mining word set by using a hidden Dirichlet distribution model to obtain the vectorized keyword set and mining word set;
acquiring vectorization similarity of the vectorization-processed keyword set and the mining word set;
and determining the vectorization similarity as the similarity between the words in the keyword set and the words in the mining word set.
5. The method of claim 3 or 4, wherein determining the mining word reserve price according to the similarity and the optimal reserve price comprises:
multiplying the similarity between at least one keyword in the keyword set and the words in the mining word set by the optimal reserve price corresponding to the keyword to obtain at least one intermediate reserve price;
accumulating the similarity between at least one keyword in the keyword set and the words in the mining word set to obtain a similarity sum value;
and determining the ratio of the sum of all the intermediate reserve prices to the similarity sum value as the reserve price of the mining word.
6. An apparatus for determining a reserve price in an advertising marketplace, the apparatus for use in a processor, comprising:
the system comprises an acquisition module, a searching module and a searching module, wherein the acquisition module is used for acquiring a keyword set and a digging word set in the historical advertisement, the keyword set comprises at least one keyword in the historical advertisement, the keyword in the keyword set is derived from a result of analyzing and processing a display log of the historical advertisement, and the digging word set is formed by the keywords dug in the historical advertisement according to a click rate;
the determining module is used for determining the optimal reserve price corresponding to the keywords in the keyword set and the reserve price of the mining words corresponding to the keywords in the mining word set;
and the processing module is used for determining a final reserve price corresponding to the keyword according to the optimal reserve price and the reserve price of the mining word and applying the final reserve price to a reserve price dictionary on the line.
7. The apparatus of claim 6, wherein the determining module is configured to:
and determining the reserve price of the mining words corresponding to the mining word set according to the keyword set, the optimal reserve price and the mining word set.
8. The apparatus of claim 7, wherein the determining module is configured to:
acquiring similarity between words in the keyword set and words in the mining word set;
and determining the reserve price of the mining words according to the similarity and the optimal reserve price.
9. The apparatus of claim 8, wherein the determining module is configured to:
vectorizing the keyword set and the mining word set by using a hidden Dirichlet distribution model to obtain a vectorized keyword set and a vectorized mining word set;
acquiring vectorization similarity of the vectorization-processed keyword set and the mining word set;
and determining the vectorization similarity as the similarity between the words in the keyword set and the words in the mining word set.
10. The apparatus of claim 8 or 9, wherein the determining module is configured to:
multiplying the similarity between at least one keyword in the keyword set and the mining word set by the optimal reserve price corresponding to the keyword to obtain at least one intermediate reserve price;
accumulating the similarity between at least one keyword in the keyword set and the words in the mined word set to obtain a similarity sum value;
and determining the ratio of the sum of all the intermediate reserve prices to the similarity sum value as the reserve price of the mining word.
11. A terminal for determining reserve prices in an advertising marketplace, comprising:
a memory;
a processor; and
a computer program;
wherein the computer program is stored in the memory and configured to be executed by the processor to implement a method of reserve price determination in an advertising market as claimed in any one of claims 1 to 5.
12. A storage medium, characterized in that the storage medium is a computer-readable storage medium having stored thereon a computer program;
the computer program is executed by a processor to implement a method of determining reserve prices in an advertising market as claimed in any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810682467.7A CN110648157B (en) | 2018-06-27 | 2018-06-27 | Method, device, terminal and storage medium for determining reserve price in advertisement market |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810682467.7A CN110648157B (en) | 2018-06-27 | 2018-06-27 | Method, device, terminal and storage medium for determining reserve price in advertisement market |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110648157A CN110648157A (en) | 2020-01-03 |
| CN110648157B true CN110648157B (en) | 2023-04-07 |
Family
ID=68988884
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810682467.7A Active CN110648157B (en) | 2018-06-27 | 2018-06-27 | Method, device, terminal and storage medium for determining reserve price in advertisement market |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110648157B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113724069B (en) * | 2021-08-31 | 2024-02-13 | 平安科技（深圳）有限公司 | Deep learning-based pricing method, device, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102479367A (en) * | 2010-11-30 | 2012-05-30 | 百度(中国)有限公司 | Method of determining reservation price of network popularization resource and device |
| CN104331823A (en) * | 2014-11-19 | 2015-02-04 | 北京奇虎科技有限公司 | Method and device for determining keyword reservation price in issued information |
| CN106033583A (en) * | 2016-05-18 | 2016-10-19 | 杭州算子科技有限公司 | Advertisement paid listing method and system applied to electronic commerce |
| KR20170014174A (en) * | 2015-07-29 | 2017-02-08 | 김영덕 | Price comparison service system based on point reserving ratio and control method thereof |
| CN108108992A (en) * | 2016-11-25 | 2018-06-01 | 百度在线网络技术（北京）有限公司 | Advertisement charge processing method and device based on entity cluster |
-
2018
- 2018-06-27 CN CN201810682467.7A patent/CN110648157B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102479367A (en) * | 2010-11-30 | 2012-05-30 | 百度(中国)有限公司 | Method of determining reservation price of network popularization resource and device |
| CN104331823A (en) * | 2014-11-19 | 2015-02-04 | 北京奇虎科技有限公司 | Method and device for determining keyword reservation price in issued information |
| KR20170014174A (en) * | 2015-07-29 | 2017-02-08 | 김영덕 | Price comparison service system based on point reserving ratio and control method thereof |
| CN106033583A (en) * | 2016-05-18 | 2016-10-19 | 杭州算子科技有限公司 | Advertisement paid listing method and system applied to electronic commerce |
| CN108108992A (en) * | 2016-11-25 | 2018-06-01 | 百度在线网络技术（北京）有限公司 | Advertisement charge processing method and device based on entity cluster |
Non-Patent Citations (1)
| Title |
|---|
| 搜索引擎营销系统关联方行为分析及其博弈机制研究;王艳红;《中国优秀硕士学位论文全文数据库经济与管理科学辑》;20130215(第2期);J152-669 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110648157A (en) | 2020-01-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11397772B2 (en) | Information search method, apparatus, and system | |
| CN107609101B (en) | Intelligent interaction method, equipment and storage medium | |
| US11238870B2 (en) | Interaction method, electronic device, and server | |
| US8615434B2 (en) | Systems and methods for automatically generating campaigns using advertising targeting information based upon affinity information obtained from an online social network | |
| US20170164049A1 (en) | Recommending method and device thereof | |
| EP3617952A1 (en) | Information search method, apparatus and system | |
| CN110830812B (en) | Similar anchor classification model training method, anchor recommendation method and related device | |
| US20170140464A1 (en) | Method and apparatus for evaluating relevance of keyword to asset price | |
| CN110475155B (en) | Live video hot state identification method, device, equipment and readable medium | |
| US20150379571A1 (en) | Systems and methods for search retargeting using directed distributed query word representations | |
| WO2018040069A1 (en) | Information recommendation system and method | |
| TW201911080A (en) | Search method, search server and search system | |
| US20190303980A1 (en) | Training and utilizing multi-phase learning models to provide digital content to client devices in a real-time digital bidding environment | |
| WO2020043001A1 (en) | Advertisement placement method, method for determining popularization population, server, and terminal | |
| JP6405704B2 (en) | Information processing apparatus, information processing method, and program | |
| CN110222256B (en) | Information recommendation method and device and information recommendation device | |
| CN109275047A (en) | Video information processing method and device, electronic equipment, storage medium | |
| CN112508612A (en) | Method for training advertisement creative generation model and method and related device for generating advertisement creative | |
| CN114168843A (en) | Search word recommendation method, device and storage medium | |
| CN110990598A (en) | Resource retrieval method and device, electronic equipment and computer-readable storage medium | |
| CN113806588A (en) | Method and device for searching video | |
| US20150278907A1 (en) | User Inactivity Aware Recommendation System | |
| CN112328889A (en) | Method and device for determining recommended search terms, readable medium and electronic equipment | |
| CN110648157B (en) | Method, device, terminal and storage medium for determining reserve price in advertisement market | |
| KR101318975B1 (en) | A Personalized Intelli gent Method for Recommending Mobile Applications |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200423 Address after: 310052 room 508, floor 5, building 4, No. 699, Wangshang Road, Changhe street, Binjiang District, Hangzhou City, Zhejiang Province Applicant after: Alibaba (China) Co.,Ltd. Address before: Unit 01, 13 Floors, B Tower, Pingyun Plaza, 163 Xiping Yun Road, Huangpu Avenue, Tianhe District, Guangzhou City, Guangdong Province Applicant before: UC MOBILE (CHINA) Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |