KR20060057722A - Method for providing a search service and system for executing the method - Google Patents

Abstract

The present invention relates to a method of providing a search service and a search system for performing the method. More particularly, the present invention relates to a method for selecting a search result based on an average click rate and a bid of each search result and a search system for performing the method. It is about.

The search service providing method according to the present invention maintains a search result database including a search term, a search result associated with the search term, a bid associated with the search result, an impression index associated with the search result, or an advertiser identifier associated with the search result. Calculating an average click rate for each of the search results based on a current time point, which is an average of click rates per unit period in a past period of time (the click rate for each unit period is the number of times the search result is provided to the user) Number of clicks, that is, the number of clicks which is the number of times the user selects the provided search result, and the number of clicks calculated for each unit period), and the impression index based on the average click rate and bid for each search result is calculated and the search result database is searched. Recording and updating in association with the results Receiving a search request including a predetermined search word from a user, selecting a predetermined number of search results having the highest exposure index among the search results associated with the search word by referring to the search result database, and the selected predetermined search result Generating a search result list including a number of search results and providing the same to the user.

Keyword ads, search terms, search results, clickthrough rates, average clickthrough rates, bids.

Description

Method for Providing A Search Service And System For Executing The Method}

1 is a diagram illustrating a network connection of a search system according to the present invention.

2 is a flowchart illustrating a search service providing method according to an embodiment of the present invention.

3 is a diagram illustrating an example of a data structure included in a search result database adopted according to an embodiment of the present invention.

4 is a table illustrating a click rate for a predetermined search result on a daily basis, for a search result associated with a predetermined search word.

FIG. 5 is a diagram showing a table showing an example of an exposure index, a bid, an advertisement execution amount, and a registration time of a search result for a predetermined search result to explain the present invention.

6 is a block diagram showing a search system according to another embodiment of the present invention.

7 is an internal block diagram of a general purpose computer device that may be employed to perform a method of providing a search service according to the present invention.

<Explanation of symbols for the main parts of the drawings>

600: search system                 

601: search result database 602: exposure index calculator

603: database management unit 604: user input unit

605: search result selection unit 606: search result list provider

The present invention relates to a method of providing a search service and a search system for performing the method. More particularly, the present invention relates to a method for selecting a search result based on an average click rate and a bid of each search result and a search system for performing the method. It is about.

A search system that provides a recent search service does not merely provide search results in response to a user's search request, but in return for providing the search results to the user, or in return for selecting the provided search results, The so-called "keyword advertising service" for charging a predetermined advertising fee is simultaneously provided to the advertiser who registered the search results.

However, in order to generate a predetermined advertising effect by exposing a search result of an advertiser to a user, the user recognizes the search result or selects the search result and links to a web page related to the advertiser. The search results may be arranged in a position and a form that can be easily recognized by the user in a search result page.                         

As the type of information searchable on the Internet is enormous, the number of search results provided in response to a user's search request increases. Accordingly, only a list of search results according to the search request, which is several to tens of pages on the web page, can be found. Cases also occur frequently. Thus, the advertiser wants his search results to be provided in a more distinctive form, above the list of search results. Hereinafter, a list of search results provided above or in a characteristic form according to a keyword search service is referred to as a "search result list".

However, when a plurality of search results registered for a keyword advertisement exist for a predetermined search word, it is impossible or inappropriate to arrange the plurality of search results all at the top or provide a characteristic form. Therefore, the search system must perform a process of selecting a search result included in a search result list by the keyword advertisement service among search results registered for the search word.

The search system disclosed in Korean Patent Laid-Open Publication No. 2002-0019042 discloses a configuration in which, when providing a keyword advertising service, the search results are sorted according to a bid registered by an advertiser and provided to a user who makes a search request. That is, the search system is provided to the user in the form that the search results with a high bid is located above the other search results. That is, according to the related art, a search system providing a keyword advertisement service selects search results included in a search result list based on a bid.

On the other hand, in a search system that does not provide a keyword advertising service, there is known a configuration that selects and provides a search word requested by a user and a search result to be provided corresponding to the search request in order of relevance or preference.

Various methods are known for measuring the relevance or preference between a search term and a search result or information provided in connection with the search result, and one of them is how many times a user selects a given search result according to a search request. How to measure depending on whether or not. That is, the search results that are frequently selected by the users may be regarded as having high relevance or preference with respect to the search word.

However, such relevance or preference is a factor that is difficult to be introduced in a keyword advertising service that charges an advertisement fee to an advertiser in exchange for being included in a search result list and provided to a user. Therefore, in the case of providing a keyword advertising service, the conventional search system has no choice but to select search results included in a search result list provided according to the keyword advertising service depending on bids.

However, when a search result provided in a search result list included in the search result does not correspond to information desired by the user or does not provide sufficient information, the user may search for a search system that provides a search service. There is a risk of their credibility being lowered.

Accordingly, NHN Co., Ltd., which is the applicant of the present invention and provides a search service including a keyword advertisement service on a web page naver (www.naver.com), selects a search result as well as a bid, that is, a predetermined A method of selecting a search result included in a search result list is further considered in consideration of the "click rate" which is the ratio of the number of times a search result is provided to users and the number of times the user selects the provided search result.

However, when the applicant experiments with the vast database accumulated in the process of providing the keyword advertising service on the web page, and when the click rate is used in the process of selecting the search result, the search result included in the search result list is very high. It has been found that severe fluctuations can occur.

That is, because the clickthrough rate is determined by the user's choice, the volatility is high over time, depending on time of day, time of day, social issues, or for no specific reason, and some search results are selected during this cycle. Unlike the prior art in which only bids are considered, such as not selected in the next cycle, it is difficult to predict which search results are included in the search result list.

If you don't solve this problem, advertisers can't predict how their search results will be selected by adjusting their bids, making it difficult for them to decide how much they should bid, and users will receive different results each time they search. It will be difficult to find the search results again.

Accordingly, the applicant of the present invention, in order to select the search results included in the search result list, together with the click rate to provide users with a bid corresponding to the purpose of the keyword advertising service and the search results recognized as relevant to the search term Reflecting, however, it has led to the development of a search service providing method and a search system for performing the method that can solve the problem that the search results included in the search result list is severely changed according to the change in the click rate.

The present invention has been made to improve the prior art as described above, and provides a search service providing method and a search system that can select a search result in consideration of the bid associated with the search results, the average click rate calculated for the search results together It aims to provide.

In addition, the present invention is to provide a search service providing method and a search system that can ensure the quality of the search results provided by the keyword advertising service by selecting the search results based on the bid and the average click rate for another object do.

Another object of the present invention is to provide a search service providing method and a search system that allow an advertiser to predict which search results can be selected by selecting a search result using an average click rate whose variation is smaller than that of the click rate. It is done.

In addition, the present invention determines the M value according to the number of inputs of the search word in order to calculate the average click rate, and calculates the average click rate by reflecting only the click rate during the unit period having the number of impressions (the number of times the search results are provided) equal to or greater than the M value. It is another object of the present invention to provide a search service providing method and a search system that can further ensure that the fluctuation range is small.

In order to achieve the above object and solve the above-mentioned problems of the prior art, the search service providing method according to the present invention, a search word, a search result associated with the search word, a bid associated with the search result, an exposure associated with the search result Maintaining a search results database including an index or an identifier of an advertiser associated with the search result, and calculating an average click rate, which is the average of the click rate per unit period over a period of time, based on the current time point, for each of the search results ( The click rate for each unit period is a number of times calculated for each unit period by the ratio of the number of impressions that are the number of times the search result is provided to the user and the number of clicks that are the number of times the user selects the provided search result. Calculate the impression index based on average clickthrough rate and bid; Recording and updating the result in association with the search result in a result database; receiving a search request including a predetermined search word from a user; wherein the exposure index is higher among search results associated with the search word by referring to the search result database; Selecting a predetermined number of search results, and generating and providing a search result list including the selected predetermined number of search results to the user.

According to one aspect of the invention, the impression index is determined as a product of the average click rate and the bid.

In addition, according to another aspect of the present invention, the exposure index, calculating a weighted average click rate by adding a first weight value to the average click rate, calculating a weighted bid by adding a second weight value to the bid, the weight Calculating a multiplication value of an average click rate and the weighted bid, and determining the calculated multiplication value as an exposure index.

According to yet another aspect of the present invention, the predetermined time period is determined as a sum of N unit periods, and the calculating of the average click rate during the predetermined time period from the present time point for each of the search results may include: For each of the search results, identifying N unit periods in which the number of impressions in the unit period is M or more in the order of closest to the current point of time, and for each of the identified N unit periods, the click rate for each unit period is summed. And calculating the average click rate during the predetermined period by dividing the sum of the click rates by the predetermined period.

Further, according to another aspect of the present invention, the step of identifying each of the N unit periods in the order in which the number of impressions during the unit period of time M or more for each of the search results in the order close to the current time point, the user during the last second period Calculating an average for each unit period of an input number that is a number of times a search word associated with the search result is inputted, and determining the M value to be proportional to the calculated average for each unit period.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a network connection of a search system according to the present invention.

The search system 100 maintains a search result for each search term in a predetermined search result database, and when a search request for a search term is input from the user terminal 110, the search system 100 includes a search result list including search results corresponding to the search term. Transmit to terminal 110.                     

The user may input a search request to the search system 100 using the user terminal 110 and receive a list of search results in response to the search request. The user terminal 110 may include a code division multiplexing access (CDMA) module, a Bluetooth module, an infrared communication module (IrDA), a wired / wireless device such as a personal digital assistant (PDA), a smart phone, a handheld PC, a mobile phone, and the like. By providing a predetermined communication module such as a LAN card, a communication terminal capable of connecting to the search system 100 through a wired or wireless internet network is collectively referred to.

According to the present invention, the search system 100 searches for a plurality of searches corresponding to the search terms according to an impression index calculated based on 1) a bid suggested by an advertiser for each search result of the search term, and 2) an average click rate. It decides which search results from among the results are provided or whether the search results are placed on the first screen of the search results web page.

The bid means the maximum cost that the advertiser is willing to pay as an advertisement fee when the advertiser's search results are provided to the user in association with a predetermined search term and the user selects the search results. The specific meaning of average click rate is mentioned later in each Example.

Hereinafter, a search service providing method according to an embodiment of the present invention will be described with reference to FIG. 2. 2 is a flowchart illustrating a search service providing method according to the present embodiment.

In operation 201, the search system 100 generates a search result database including a search term, a search result associated with the search term, a bid associated with the search result, an impression index associated with the search result, or an advertiser identifier associated with the search result. Keep it. 3 is a diagram illustrating an example of a data structure included in the search result database. Each data included in the search result database is updated regularly or irregularly. For example, when a new advertiser registers a new search result for the search word, a search result associated with the search word is additionally recorded. The search result is information provided to the user according to a search request, and may be used as information for providing a link to the web page including link information to a predetermined web page.

In step 202, the search system 100 calculates an average click rate for each of the search results. The average clickthrough rate is an average of clickthrough rates per unit period for a certain period in the past, based on the current time point. Hereinafter, a case in which the unit period is selected as one day will be described as an example. However, this is merely exemplary, and the unit period may be arbitrarily selected by the operator of the search system 100 in various manners such as 1 hour, 10 hours, 2 days, and so on.

Click rate can be calculated by dividing the number of clicks by the number of impressions. The number of impressions means the number of times a predetermined search result is provided to the user according to the keyword advertisement search, and the number of clicks means the number of times the provided search result is selected by the user.

The average clickthrough rate may be calculated differently depending on how the predetermined period is selected. A process of calculating the average click rate will now be described with reference to FIG. 4.

(1) When calculating average click rate according to the first method

Table 1 of FIG. 4A illustrates a click rate for a predetermined search result on a daily basis, for a search result associated with a predetermined search word. For example, Table 1 calculates the daily click rate for the search result 301 in association with the search term "flower delivery."

The first method selects a continuous period of time in the past from the present time point as the above-described constant period. Referring to Table 1, if the current point in time is October 10, 2004 at 10 o'clock, several consecutive days before the current point in time, for example, five days from 0 o'clock October 5, 2004 to 24 o'clock October 9, 2004 Consecutive periods of time can be determined as said constant period.

Therefore, from October 5, 2004 to October 9, 2004, the average click rate can be calculated by adding all of the daily click rates and dividing by 5 days. Referring to Table 1, the average clickthrough rate for the search result 301 is as follows.

(2) When calculating average click rate according to the second method

The second method selects the sum of predetermined N unit periods as the predetermined period.

The N unit periods, that is, N days, may not be continuous unlike in the first method. According to the second method, the retrieval system 100 identifies N, for example, five, in a close order, based on the current time point, on a unit period of time, i.e., the number of impressions for one day or more.                     

If the M value is 20, the days on which 20 or more clicks are input are indicated by hatching in Table 2 of FIG. 4B. For example, October 8, 2004 where 80 clicks were entered.

The search system 100 may identify five past days in which 20 or more clicks have been input in the reverse order of the time flow based on the current time point (10 October 31, 2004). The search system 100 may calculate the average click rate by summing the click rates of the identified days and dividing by 5, respectively. Referring to Table 2, the average click rate of the search result 301 is as follows.

Meanwhile, in the present embodiment, for the convenience of explanation, the average click rate for 5 days, which is a relatively short period, was calculated. However, the search system 100 determines how long the average click rate is to be calculated, that is, to select the predetermined period. ) Can be determined using empirical and statistical methods. For example, the search system 100 may calculate an average clickthrough rate over the last 30 days based on the current time.

In addition, according to another embodiment of the present invention, the M may be determined differently for each search word by reflecting the number of times the search word is input. The number of times the user inputs the search word, which is the number of times the user inputs the search word, may reflect the popularity of the search word. For example, the search system 100 calculates an average for each unit period of the number of inputs (that is, the number of times a search request is made while entering the search term), which is the number of times the search term has been input in the past based on the current time point. The M value may be selected in proportion to the average of each unit period. In order to distinguish between the predetermined period selected for calculating the average click rate and the predetermined period selected for calculating an average for each unit period of the number of inputs of the search word, the latter is referred to as a “second period”. The second period may be selected by an operator of the search system 100, and may be selected in consideration of the predetermined period (N days) selected to calculate the average click rate.

30 = 100으로 산출된다. For example, when the second period is set to 30 days, if the search term "flower delivery" has been input from the user in the last 30 days, a total of 3,000 times, the average for each unit period (day) of the input number of the search term "flower delivery" Silver 3000

Calculated as 30 = 100.

The search system 100 may calculate the M value for each search word in proportion to the daily average for 30 days of the number of times the search term “flower delivery” is input. For example, M may be calculated as follows.

M = associated with the term "flower delivery"

5% of the daily average of the number of entries in 30 days for the query "flower delivery" =

100 × 20% = 20

That is, the search system 100 may calculate the average click rate only for a day of 20 or more impressions for a search result associated with the search word "flower delivery".

Meanwhile, in the above embodiment, 20% of the daily average is selected as M, but 10% of the daily average is selected as M, such that the daily average for each unit period for a predetermined period for each keyword is reflected to the M value to some extent. The operator of search system 100 can determine statistically and empirically.

The average click rate calculated according to the first method or the second method is a factor reflected in the process of selecting a search result for the search word as described below.

In addition, as described above, the average click rate is based on "the number of impressions that are the number of times that a user is provided with a search result in connection with a predetermined search term and the click rate that is the ratio of the number of clicks that the user clicks on a predetermined search result". As calculated on the basis of the results, the average click rate reflects the user's preference or the relevance between the search term and the search result. In addition, since the average click rate has a characteristic of a moving average of the click rate for the last few days from the present time point, the average click rate calculated every time according to a predetermined search result has a smaller variation than the daily click rate.

Using Table 2 of FIG. 4B as an example, if search results 301 had 70 impressions on October 10 and a clickthrough rate of 12, the average clickthrough rate over the past N days (ie, 5 days) on October 11 If calculated according to the second method, the newly calculated average click rate is as follows.

Thus, compared to the previously calculated average clickthrough rate of 15.3, the difference from the newly calculated average clickthrough rate is only -1. On the other hand, if you look at the number of clicks per day, compared to 15 times on October 9, two days ago, the click rate on October 10, the day before, is 12, the difference is three.

Therefore, when selecting the search result by reflecting the average click rate, the possibility of variation of the selected search result is lower than that of simply selecting the search result by reflecting the click rate during the unit period such as the daily click rate.

Meanwhile, the search system 100 may periodically calculate the average click rate, such as calculating the average click rate at 2:00 am every day.

In operation 203, the search system 100 calculates an exposure index for each search result based on the average click rate calculated for the search result and the bid associated with the search result as described above.

1) According to an embodiment of the present invention, the exposure index is calculated as follows.

Impression index = average clickthrough rate × bid

2) In addition, according to another embodiment of the present invention, the exposure index is calculated as follows.

Impression index = weighted average clickthrough rate × weighted bid

= (Average clickthrough rate + a) × (bid price + b)                     

A and b are respectively a first weight value added to the average click rate and a second weight value added to the bid. Depending on how the first weight value and the second weight value are selected, the operator of the search system 100 may determine whether to further reflect preference or relevance or to further reflect bids.

The operator of the retrieval system 100 may appropriately select a and b according to statistical or empirical methods.

The exposure index according to the present embodiment is calculated based on i) for each search result, ii) for each search result, and iii) for each click result. Referring to FIG. 3, the average click rate of the search result 301 associated with the search word “flower delivery” is 15.3, and the bid amount suggested by the advertiser who registered the search result 301 is 950 won. Thus, the impression index for the search result 301 associated with the search term "flower delivery" is calculated as 15.3 x 950 = 14535 from the average click rate x bid.

Therefore, those skilled in the art can know that the impression index is changed when the advertiser registers by changing the bid, or when the average clickthrough rate is changed. That is, since the impression index is calculated based on the average click rate, the impression index reflects how much users prefer the search result 301.

In addition, since the advertiser may register the same search result 301 for each of the search terms "flower delivery", "garland", and the like, the exposure index is preferably calculated for each search word and for each search result.

In addition, the search system 100 may update the exposure index in real time whenever the bid is changed or the average click rate is changed for a predetermined search result associated with a predetermined search word. In addition, the search system 100 may periodically calculate the exposure index in consideration of the system load.

In step 204, the search system 100 records and updates the exposure index calculated for each search result in the search result database in association with the search result. The search result database of FIG. 3 illustrates a case in which the impression index 305 is calculated by multiplying the average click rate 304 and the bid 303.

In step 205, the user sends a search request for a predetermined search term to the search system 100. For example, the search term is "flower delivery."

In operation 206, the search system 100 selects a predetermined number of search results having a high exposure index among search results associated with the search word by referring to the search result database. For example, the search system 100 may select five search results with a high exposure index.

For example, if the exposure index of the search results associated with the search term "flower delivery" is high in order, respectively, 19000, 18000, 17000, 16500, 16200 (for search result 302), 14535 (search result 301). ), 14000,... , Then select five search results with an exposure index of 19000, 18000, 17000, 16000, 16200. Therefore, the search result 301 having the sixth highest exposure index as 14535 is not included in the five search results.

In addition, according to another embodiment of the present invention, the search system 100, if there is a search result with the same exposure index is not able to select the top five, the following priority is given to only five search results each time Can be selected. Table 3 of FIG. 5 is a table showing an example of an impression index, a bid, an advertisement execution amount, and a registration time of the search result, for each search result associated with the search word "flower delivery". Table 3 shows a case where the search results are sorted according to the exposure index.

Because of the search results 4 through 7 currently having an exposure index of 16500, it is not easy for the search system 100 to select only the top five search results. Therefore, the search system 100 according to the present embodiment gives priority to search results having a high bid when the impression index is the same.

Referring to Table 3, search result 7 has a smaller bid than search results 4 to 6 having the same exposure index. Therefore, the search system 100 does not select the search result 7.

On the other hand, since the search results 4 to 6 have the same priority even after the impression index is further considered, it is not easy for the search system 100 to select only five search results again. Therefore, according to another exemplary embodiment of the present invention, the search system 100 selects a search result by further considering an advertisement execution amount for a search result having the same impression index and a bid.

As described above, the bid may mean a maximum advertising cost that the advertiser can pay, and the actual advertising fee charged to the advertiser may be different from the bid. In this specification, an advertisement fee charged to the advertiser when users select a search result is called an advertisement execution cost. For example, the search system 100 may determine the bid amount as the advertisement execution amount, or may determine the amount of the advertisement cost by adding a predetermined amount to the bid of the search result corresponding to the next rank of the search result.

In addition, according to another embodiment of the present invention, if five search results cannot be selected even though the impression index, the bid amount, and the advertisement execution amount are considered, the search result registered first may be selected by further considering the registration date of the search results. .

Accordingly, the search system 100 may include i) an impression index, ii) a bid for search results with the same impression index, iii) an ad execution cost for a search result with the same impression index and bid, iv) an impression index, a bid, and an ad flight. For search results having the same cost, a predetermined number of search results may be selected in consideration of the search result registration time.

Referring to Table 3 of FIG. 5, only the search results 4 to 6 are selected in consideration of the bid for the search results 4 to 7 having the same impression index, and the ad execution amount for the search results 4 to 6 having the same impression index and the bid. Consider further, and select the search results 4 and 5. In consideration of the advertisement execution amount, since all five search results are selected, such as the search results 1 to 5, in the case of Table 3, the search system 100 does not need to consider the registration timing of the search results further.

In step 207, the search system 100 generates a search result list including the selected five search results and provides them to the user. That is, the search system 100 transmits the search result list to the user terminal 110.

Still referring to FIG. 3, when comparing the search result 301 and the search result 302, a search result 302 having a lower bid but a large average clickthrough rate is different from the search result 301 which has a higher bid. ) Is selected as the search result included in the search result list. Thus, according to the present embodiment, an advertiser may, i) raise a bid, ii) present a bid to some extent in order for his or her search result to be included in a search result list and be provided to users. It is possible to adopt a method that allows users to choose a lot by organizing in a form and content that they may prefer. Since the bid is used as a criterion for calculating an advertisement execution cost, which is an advertisement fee charged to the advertiser when each search result is clicked by a user, it is difficult for the advertiser to raise the bid without risk.

Thus, according to the present invention, i) by reflecting the click-through rate for the search results when selecting the search results, so that the user's preferred search results can be provided to users more, ii) not the click-through rate for the search results themselves, By using the average value of clickthrough rates per unit time period over the past period of time, we ensure that the search results provided each time a search request does not fluctuate significantly, and iii) advertisers want to increase their chances of selecting their search results, The quality of the search results is also taken into account. Therefore, according to the present invention, the quality of the search results provided by the search system 100 is increased, thereby increasing the user's preference for the search service itself provided by the search system 100. That is, users prefer a search service provided by the search system 100 to other search services.

In addition, advertisers who find that their search results are not preferred by their users, in terms of average clickthrough rates, can increase their bids by lowering their average clickthrough rates so that they can still be included in the search results list. Therefore, it is possible to increase the rate of return in the search system 100. In particular, according to the present invention, since the variation in the average click rate is small, the advertiser can roughly predict the degree to which the search results can be selected by selecting the bid.

Meanwhile, in the above-described embodiment, when a search request is input in step 205, the search result is selected in step 206. However, this is merely an example, and the search system 100 displays an impression index for each search word. May select the top five search results in real time or periodically in advance, and provide the five preselected search results when a user requests a search.

In addition, the search system 100 may generate the search result list to arrange the five search results included in the search result list in the order of exposure. In addition, the search system 100 may generate the search result list such that five search results are arranged according to a bid. In addition, each time the search system 100 provides the user with a list of search results, the search system 100 may allow the five search results to be newly sorted in a rolling manner. However, the present invention relates to a method of selecting a predetermined number of search results included in a search result list, and the scope of the present invention is provided in which order, or in what form, the selected search results are provided in a sorted order. Not limited by

In addition, embodiments of the present invention include computer-readable media containing program instructions for performing various computer-implemented operations. The program recorded on the computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those having skill in the computer software arts.

Hereinafter, another search system will be described for another embodiment of the present invention. 6 is a block diagram showing a search system according to the present embodiment.

The search system 600 includes a search result database 601, an exposure index calculator 602, a database manager 603, a user input unit 604, a search result selector 605, and a search result list provider 606. It includes.

The search results database 601 includes a search term, a search result associated with the search term, a bid associated with the search result, an impression index associated with the search result, or an advertiser identifier associated with the search result. As described later, the exposure index for each search result included in the search result database 601 is updated in real time or periodically every time the exposure index calculator 602 calculates the exposure index.

The exposure index calculator 602 calculates, for each unit period, an impression number, which is a number of times the search result is provided to the user for each search result associated with each search word. For example, when a day is selected as the unit period, the number of times the search result is provided to the user from 0 o'clock of the day to 12 o'clock of the day is calculated.

In addition, when the search result is provided to the user for each search result associated with each search word, the exposure index calculator 602 calculates the number of clicks, which is the number of times the user selects the search result every unit period, for example, every day. .

단위 기간 별 노출수)로 산출될 수 있다. In addition, the exposure index calculator 603 calculates a click rate for each unit period from the number of impressions per unit period and the number of clicks for each unit period calculated for each search result in association with each search word. Clickthrough rate per unit period is (clicks per unit period)

Impressions per unit period).

Next, the exposure index calculator 602 calculates an average click rate, which is an average of click rates for a predetermined period of time, based on the current time. For example, the impression index calculator 602 may calculate an average of click rates for the last 30 days. The predetermined period is defined as N unit periods, and when a day is selected as a unit period as in this embodiment, the predetermined period corresponds to N days. Hereinafter, the case where 30 days are selected as N days is taken as an example.

In this case, as in the above-described embodiment, the average click-through rate may be calculated by summing all the click-through rates for 30 consecutive days and dividing by 30. The click-through rate of the selected day is selected by selecting 30 days only on days that satisfy a predetermined condition. You can also calculate the average clickthrough rate by summing all the numbers and dividing by 30.

More specifically, the exposure index calculator 602 selects 30 days in order of closeness based on the current time point among the number of impressions for one day or more of the search result.

In addition, according to another embodiment of the present invention, the exposure index calculator 602 may set the M value for each search word in consideration of the number of inputs of the search word. For example, the search system 100 may calculate an average for each unit period of the number of inputs (that is, the number of times a search request is made while entering the search word), which is the number of times the search word has been input in the past second period based on the current time point. That is, the daily average may be calculated and the M value may be selected in proportion to the average. The second period may be selected by an operator of the search system 100, and may be selected in consideration of a predetermined period (N days) selected to calculate the average click rate.

The impression index calculator 602 calculates an impression index for the search result based on the average click rate calculated as described above and a bid registered by the advertiser of the search result. For example, the impression index calculator 602 may calculate an impression index by multiplying an average click rate and a bid. In addition, the exposure index calculator 602 may calculate the exposure index by multiplying the weighted average click rate obtained by adding the first weighted value to the average click rate and the weighted bid obtained by adding the second weighted value to the bid.

When the exposure index is calculated by the exposure index calculator 602, the database manager 603 records the exposure index in real time or periodically in association with the search result in the search result database 601, thereby searching the search result database ( 601) is updated.

The user input unit 604 receives a search request including a predetermined search word from the user. The user may access the search system 600 using a communication terminal such as a PC and input a search request.                     

The search result selector 605 refers to the search result database 601 and selects a predetermined number of search results of which the exposure index is higher among search results associated with the search word. The search result selector 605 may select a predetermined number of search results based on the exposure index for each search term, regardless of the input time of the search request.

The search result list providing unit 606 generates a search result list including the selected predetermined number of search results and provides them to the user.

7 is an internal block diagram of a general purpose computer device that may be employed to perform a method of providing a search service according to the present invention.

Computer device 700 includes one or more processors 710 connected to a main memory device including random access memory (RAM) 720 and read only memory (ROM) 730. The processor 710 is also called a central processing unit (CPU). As is well known in the art, the ROM 730 serves to transfer data and instructions to the CPU unidirectionally, and the RAM 720 typically transfers data and instructions bidirectionally. Used to. RAM 720 and ROM 730 may include any suitable form of computer readable media. Mass storage 740 is bidirectionally coupled to processor 710 to provide additional data storage capabilities, and may be any of the computer readable recording media described above. The mass storage device 740 is used to store programs, data, and the like, and is typically an auxiliary memory device such as a hard disk which is slower than the main memory device. Certain mass storage devices such as CD ROM 760 may be used. The processor 710 may include one or more input / output interfaces, such as video monitors, trackballs, mice, keyboards, microphones, touchscreen displays, card readers, magnetic or paper tape readers, voice or handwriting readers, joysticks, or other known computer input / output devices. 750 is connected. Finally, the processor 710 may be connected to a wired or wireless communication network through the network interface 770. Through this network connection, the procedure of the method described above can be performed. The apparatus and tools described above are well known to those skilled in the computer hardware and software arts.

The hardware device described above may be configured to operate as one or more software modules to perform the operations of the present invention.

While specific embodiments of the present invention have been described so far, various modifications are possible without departing from the scope of the present invention.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined not only by the claims below, but also by the equivalents of the claims.

As described above, the present invention has been described by way of limited embodiments and drawings, but the present invention is not limited to the above-described embodiments, which can be variously modified and modified by those skilled in the art to which the present invention pertains. Modifications are possible. Accordingly, the spirit of the invention should be understood only by the claims set forth below, and all equivalent or equivalent modifications thereof shall be said to belong to the scope of the invention.

According to the present invention, a method and a search service providing method for selecting a search result in consideration of a bid associated with a search result and an average click rate calculated for the search result, which are devised to improve the conventional technology as described above. A system is provided.

In addition, according to the present invention, by selecting a search result based on the bid and the average click rate, there is provided a search service providing method and a search system that can ensure the quality of the search results provided according to the keyword advertising service.

In addition, according to the present invention, a search service providing method and a search system are provided that allow an advertiser to predict which search results can be selected by selecting a search result using an average click rate whose variation is smaller than the click rate.

In addition, according to the present invention, in order to calculate the average click rate, the M value is determined according to the number of inputs of the search word, and the average click rate is reflected only by the click rate for a unit period having the number of impressions or more (the number of times the search result is provided). There is provided a search service providing method and a search system that can further ensure that the variation is small by calculating.

Claims (1)

In the search service providing method in a search system, Maintaining a search results database comprising a search term, a search result associated with the search term, a bid associated with the search result, or an advertiser identifier associated with the search result; Receiving a search request including a predetermined search word from a user; Selecting a predetermined number of search results having the highest bid among search results associated with the search term by referring to the search result database; And Generating a search result list including the selected predetermined number of search results and providing the same to the user; Search service providing method comprising a.

Images