JP2010079625A - Search device, search system, and search method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a search device, a search system and a search method, for obtaining information that emphasizes seasonality or trend. <P>SOLUTION: The search system 200 is provided with: a variation extraction part 201 for analyzing time-series keyword data to extract a seasonal variation and a trend variation; a seasonal keyword information storage part 202 for accumulating the seasonal variation extracted by the variation extraction part 201, a trend variation, and a keyword in association with one another; a search result storage part 203 for storing a result of input keyword search using the input keyword set by a user using a search engine 206; and a re-evaluation part 204 for using weight to the seasonal variation and trend variation set by the user to weight the seasonal variation of the seasonal keyword or the trend variation included in the result stored in the search result storage part 203, and rearranging results of the keyword search on the basis of the result. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a search device, a search system, and a search method, and more particularly to a search device, a search system, and a search method that place importance on the seasonality or trend of a keyword included in a search keyword or search target sentence.

  2. Description of the Related Art Conventionally, in a search by keyword, a search engine is used that extracts information including the search keyword in the information by inputting the keyword on the search screen and displays the search result in order of the information with the highest relevance. . Furthermore, a search device has been proposed that automatically presents recommended document information by taking into account the number of appearances of the latest keyword information using the latest keyword information extracted from the search results (for example, Patent Document 1). ). However, in such a system, the latest keyword includes a keyword having seasonality, and only information including seasonality is searched. In addition, in a search device using ordinary search keywords, a search using keywords with a strong summer image, for example, keywords such as fireworks, bathing, or swimming pools, search for a plurality of information related to summer, and use keywords such as New Year, skiing, hot pot, and Christmas. In the search, a plurality of information related to winter is searched.

  On the other hand, a system has been proposed that presents information that takes into account the seasonality of keywords by giving each keyword a seasonal feeling such as the season or shipping time in advance (for example, Patent Documents 2 and 3). .

  In addition, a system for detecting an abnormal value included in sales performance data in demand forecast of seasonal products using time series information of sales has been proposed (for example, Patent Document 4).

JP-A-7-85103 JP 2003-76898 A JP 2002-304398 A JP 7-56890 A

  However, according to the conventional technology, there is a problem that there is no system for obtaining information that emphasizes seasonality even though attention can be paid to information including seasonality. In addition, only information that includes seasonality is focused on, and it is not important to focus on medium-term and long-term trends that are not constrained by seasonality, such as the extraction of quiet trends or movements that progress gradually. There is a problem of missing important information that is not done. That is, as in Patent Document 1, information that reflects the search demand at the time of search shows superior seasonality at the time of search, but information including a quiet trend or movement that gradually progresses is lower. There is a risk of becoming information and being buried in information including seasonality. In Patent Documents 2 and 3, only the seasonality or seasonal feeling is emphasized, and information that is buried in seasonality is not noted. Further, Patent Document 4 is a means for detecting an abnormal value in consideration of the trend variation and irregular variation included in the sales record data, and cannot detect information hidden in seasonality.

  As described above, the conventional search engine can perform only a keyword search, and a search that takes into account the importance of a seasonal factor or a trend factor requested by a user for a keyword has not been made. For this reason, a search result that matches the user's sensibility may not be obtained, and information requested by the user through a keyword search may not be reached quickly. In a vast information society, the fact that the information requested by the user cannot be obtained quickly is a loss of time, and there is a problem that the user cannot be satisfied.

  Therefore, the present invention has been made in view of such a problem, and the object thereof is based on the result of keyword search by a search engine or the like, and the keyword of seasonality included in an information search result including a search keyword. A search device capable of obtaining information emphasizing seasonality or information emphasizing trendiness by separating temporal fluctuations into seasonality factors and trend factors and specifying the importance of each factor by the user It is to provide.

  In order to solve the above problems, according to an aspect of the present invention, a time series data of a keyword is analyzed to extract a seasonal fluctuation amount and a trend fluctuation amount, and a season extracted by the fluctuation extraction unit A seasonal keyword information storage unit for accumulating the fluctuation amount and trend fluctuation amount in association with the keyword, and an input keyword search result (hereinafter referred to as an input keyword search result) using at least one input keyword set by the user. A search result storage unit to be stored and a set of keywords (hereinafter referred to as seasonal keywords) stored in the seasonal keyword information storage unit appear in the input keyword search result (in some cases, the frequency of elements). (Hereinafter, this keyword set is referred to as a search result seasonality keyword), and the amount of seasonal variation set by the user Using the weight for at least one of the trend fluctuation amounts, the seasonal fluctuation amount or trend of the seasonal keyword accumulated in the seasonal keyword information storage unit with respect to the search result seasonal keyword stored in the search result storage unit There is provided a search device including a re-evaluation unit that calculates a weighted evaluation index corresponding to a variation amount and rearranges the order of search results using the index.

  In such a configuration, the input keyword search result is re-searched by weighting the seasonal fluctuation amount and the trend fluctuation amount obtained by analyzing the time series frequency information of the seasonal keyword by the weight set by the user. . The weight set by the user presents the keyword having the fluctuation amount requested by the user to the user more preferentially among the seasonal fluctuation amount or the trend fluctuation amount of the search result seasonal keyword included in the input keyword search result. To function. For this reason, the information search which emphasized the seasonality or the trend which is contained in a search result seasonality keyword is attained. Therefore, information can be easily searched again according to what information the user needs.

The search result seasonal keyword obtained by the re-evaluation unit may allow all the seasonal keywords accumulated in the seasonal keyword information storage unit as described above, or an input keyword including the input keyword You may limit to seasonal keywords related to.
Relevance to the input keyword is determined by general methods such as using a synonym dictionary (thesaurus), or using a method that uses a high percentage of both characters such as “Nabe” and “Yosenabe”. can do.

  The input keyword search result stored in the search result storage unit includes a search score, and the reevaluation unit calculates the seasonal variation amount and trend variation amount of the seasonal keyword accumulated in the seasonal keyword information storage unit from the search score, and The re-evaluation search score may be calculated using the weight of the seasonal variation amount and the trend variation amount set by the user.

  The weighting of the seasonal variation amount and the trend variation amount set by the user may include a negative weighting.

The search score may be a value obtained by searching the input keyword, a reciprocal value corresponding to the search order, or a value obtained by switching the upper and lower ranks of the search order.
You may further provide the display part which displays the search result by a reevaluation part.

In order to solve the above problems, according to another aspect of the present invention, a search engine that performs an input keyword search using at least one input keyword set by a user, and a result of the input keyword search by the search engine are stored. Search result storage unit, analysis of keyword time-series data to extract seasonal variation and trend variation, and relationship between seasonal variation and trend variation extracted by variation extraction unit and keyword And using the weighting for at least one of the seasonal variation amount and the trend variation amount set by the user, and the input keyword search result stored in the search result storage unit. The seasonal variation amount or trend of the seasonal keyword stored again in the seasonal keyword information storage unit A search comprising: a re-evaluation unit that performs a search for seasonal keywords weighted by variation, and rearranges input keyword search results stored in a search result storage unit based on the weighted seasonal keyword search results A system is provided.

  In order to solve the above problems, according to another aspect of the present invention, a step of analyzing time series data of seasonal keywords and extracting a seasonal variation amount and a trend variation amount, and a season extracted by the extraction step A step of accumulating sex keyword information, a search step of executing an information search by an input keyword and capturing a search result (input keyword search result), and at least one of a seasonal fluctuation amount and a trend fluctuation amount set by a user A step of capturing weights, a step of calculating a reevaluation search score using weighting and seasonal keyword information set by the user in the search result (input keyword search result), and the re-evaluation search score based on the value of the reevaluation search score. Reordering the evaluation search scores.

  You may further provide the display step which displays the search result by a reevaluation step.

  As described above, according to the present invention, information emphasizing seasonality or information emphasizing trendiness can be obtained from a search result of an input keyword search by a search engine or the like in response to a user request.

  Exemplary embodiments of the present invention will be described below in detail with reference to the accompanying drawings. In addition, in this specification and drawing, about the component which has the substantially same function structure, duplication description is abbreviate | omitted by attaching | subjecting the same code | symbol.

Further, the “best mode for carrying out the invention” will be described according to the following item order.
(1) Background to one embodiment (2) Seasonality and trend (3) Outline of search device according to one embodiment (4) Re-evaluation search score calculation method (5) Hardware configuration (6) One implementation Flow of display processing according to embodiment (7) Flow of reevaluation processing according to one embodiment

Background to the Embodiment Conventionally, when a user inputs a search keyword, for example, information including the search keyword is extracted from information stored in a database provided by an Internet service provider, and access to each information is performed. Search engines in which information is ranked based on the number or the number of links and the search results are displayed are widely used.

  In such a search engine, the search results are based on the condition setting of each search engine, and the rankings are not always given with an emphasis on seasonality or trendiness reflecting the user's request. For this reason, when information that is far from the information that the user wants to search from the search keyword is displayed in the upper search results, it may take a long time to search for the information that the user needs, or information that is useful to the user It is difficult to search. In addition, in a keyword in which the season is imagined from the search keyword itself, a search result that emphasizes the seasonality at the time of search is displayed, and it may be difficult to obtain information that is independent of the season.

  Therefore, in the present embodiment, by analyzing the time-series frequency information of keywords, a search device that can obtain information that emphasizes seasonality or information that emphasizes trendiness based on the search result of seasonal keyword search. Will be described.

(2) Seasonality and trendiness Here, seasonality and trendiness will be described. FIG. 1 is a diagram for explaining time-series data, in which carbon dioxide (CO ₂ ) emissions in a certain area are plotted with respect to a time axis.

As shown in FIG. 1 (a), CO ₂ emissions tend to repeat the period when emissions are low and high throughout the year, including the factors associated with seasonal changes and clearly showing periodicity. FIG. On the other hand, time-series data can be separated into trend fluctuations, seasonal fluctuations (periodic fluctuations), and irregular fluctuations. Trend variation is a long-term continuous change, and seasonal variation (periodic variation) is seasonal or periodic variation. Irregular fluctuations are changes due to other factors.

Data containing seasonal factors obscure the movement of time-series data. For this reason, in order to see the original movement of time-series data, a method of removing seasonal factors is used, which is generally called a seasonal adjustment method. In the seasonal adjustment method, if the seasonal cycle is known in the time series data, the amount of fluctuation that is shifted by the seasonal period compared to the previous year or the previous month is taken as the magnitude of the trend fluctuation, and the other fluctuations are seasonal and irregular fluctuations. It is a method of calculating the trend fluctuation amount, the seasonal fluctuation amount, and the irregular fluctuation amount on the assumption that Methods widely used as seasonal adjustment methods include the Census Bureau method developed by the Census Bureau of the US Department of Commerce, the EPA (Economic Planning Agency) method developed by the Economic Planning Agency, and the MITI method developed by the Ministry of International Trade and Industry. In the marketing field, a method such as a Sarima (Seasonal Auto Regressive Moving Average) model is used for analyzing time series data. By appropriately using these methods, it is possible to easily extract the seasonal variation amount and the trend variation amount included in the time-series data of the keyword.

Fig. 1 shows the time series of trend fluctuations (trend fluctuation components), the time series of seasonal fluctuations (seasonal fluctuation components), and the time series of irregular fluctuations (irregular fluctuation components) extracted by these methods. ing. If this component can be calculated, the CO ₂ trend fluctuation amount, seasonal fluctuation amount, and irregular fluctuation amount (residual) can be specifically set.
As an example of a specific setting method, if the average power of the trend fluctuation component, the average power of the seasonal fluctuation component, and the average power of the irregular fluctuation component are TP, SP and EP, respectively, the trend fluctuation amount Mt and seasonal variation amount Ms are obtained by the following equations 1 and 2, respectively.

In the present embodiment, attention is paid to the seasonal fluctuation amount that is a seasonal change and the trend fluctuation amount that is a long-term continuous change among the fluctuation elements included in the time series data. You may pay attention to the irregular fluctuation amount which is a change by various factors, such as observation error, according to the information to perform.

(3) Overview of Search Device According to One Embodiment FIG. 2 is a diagram illustrating a functional configuration including a search system 200 according to one embodiment of the present invention. As shown in FIG. 2, the search system 200 includes a fluctuation extraction unit 201, a seasonal keyword information storage unit 202, a search engine 206, a search result storage unit 203, a reevaluation unit 204, and a display unit 205. Have. The display unit 205 may be displayed on a display of the search system 200 or may be displayed on a user terminal connected to the search system 200.

  The fluctuation extraction unit 201 analyzes the time series frequency information of the keyword, and extracts the seasonal fluctuation amount Ms, the trend fluctuation amount Mt, and other fluctuation amounts. As described above, the method for extracting the seasonal variation amount Ms and the trend variation amount Mt for each keyword can be easily extracted using a seasonal adjustment method or the like.

  The seasonal keyword information storage unit 202 stores the seasonal variation amount Ms and the trend variation amount Mt of each keyword extracted by the variation extraction unit 201 in association with the keyword. The keywords stored in the seasonal keyword information storage unit 202 are referred to as seasonal keywords.

  FIG. 3 shows a specific example of the seasonal keyword, the seasonal variation amount Ms and the trend variation amount Mt accumulated in the seasonal keyword information storage unit 202. As shown in FIG. 3, the seasonal variation amount Ms of “fireworks” with a strong summer image is 100 and the trend variation amount Mt is 50, and the seasonal variation amount Ms is larger than the trend variation amount Mt. Similarly, for the keyword “nabe” having a strong winter image, the seasonal variation amount Ms is 100 and the trend variation amount Mt is 50, and the seasonal variation amount Ms is larger than the trend variation amount Mt. Further, when viewed from the seasonal variation amount Ms, the seasonal variation amount Ms of “Yosenabe” is larger than the seasonal variation amount Ms of “nabe”, and a keyword having a stronger seasonal feeling has a larger seasonal variation amount M. I understand that. On the other hand, for keywords with low seasonality, such as “pan”, the seasonal fluctuation amount Ms is 50 and the trend fluctuation amount Mt is 100, and the trend fluctuation amount Mt may exceed the value of the seasonal fluctuation amount Ms. Recognize. For a keyword such as “Tokyo” that does not include seasonality, the seasonal variation amount Ms and the trend variation amount Mt have the same value.

  Thus, the seasonal keyword information storage unit 202 stores each seasonal keyword, its seasonal variation amount Ms, and the trend variation amount Mt. When accumulating, the values of the seasonal variation Ms and the trend variation Mt may be generalized (normalized).

  In addition, the search result storage unit 203 is a search in which information is ranked for a search result of an input keyword (which may include a sentence) input by a user using a search engine 206 provided on a network, for example. Accommodates results (input keyword search results). The search result includes the search score S of each piece of information including the input keyword. The search score S is mostly provided by an existing general information search device or the like, but may be a reciprocal of the rank of the search result or a value obtained by switching the upper and lower ranks of the search rank.

  FIG. 4 shows a search result of the keyword “nabe” by a general search engine 206. Here, for ease of explanation, only four search results will be described. As shown in Fig. 4, if the search time is winter, information that imagines winter, such as seasonal hot pot and fugu hot pot, is searched for higher, and information that is not affected by seasonality such as pots and pan lids is lower. To be searched. Specifically, the search score S (1) for articles related to Yosenabe is 100, the search score S (2) for articles related to Fugu nabe is 95, the search score S (8) for articles related to Bushi nabe is 70, and an article related to pot lids Assume that the search score S (9) is 60. At this time, if the information requested by the user is related to hot pot products, the information not influenced by the seasonality will be buried in information such as hot pot dishes along the season. It is difficult to obtain from the input keyword search result.

  As described above, a keyword search using the general search engine 206 may cause the search result storage unit 203 to store a search result that does not take into account a factor requested by the user for the keyword.

  In addition, the re-evaluation unit 204 wants the user to obtain information that emphasizes seasonality or the trend characteristic with respect to the search score S (i) of each information that is the search result obtained in the search result storage unit 203. A re-evaluation search score R (i) is obtained depending on whether the emphasized information is desired. The reevaluation search score R (i) is calculated in consideration of the seasonal variation weight Ws and the trend variation weight Wt set by a GUI (Graphical User Interface) or the like. A specific method of calculating the reevaluation search score R (i) will be described in (4) Method of calculating the reevaluation search score. The calculated re-evaluation search score R (i) is rearranged based on the magnitude of the value. That is, the re-evaluation unit 204 re-evaluates the search score S (i) from the input keyword search result stored in the search result storage unit 203, calculates the re-evaluation search score R (i), and performs the re-evaluation search. The scores R (i) are rearranged. The re-evaluation unit 204 re-searches the result of the input keyword search by the general search engine 206 so that the importance of the seasonal factor or the trend factor desired by the user is taken into account. A search result satisfying can be obtained.

  Further, the display unit 205 displays the search result obtained by the reevaluation unit 204 on the display.

  The search result calculated from the search score S (i) and based on the re-evaluated search score R (i) rearranged is displayed by the display unit 205, and information satisfying the user's request is displayed at the top. Therefore, information matching the user's request can be quickly reached from the search result by the seasonal keyword search.

(4) Re-evaluation search score calculation method The re-evaluation search score R (i) is a search score S (i) of each information included in the search result by the input keyword search so that the information requested by the user is searched. In addition, the weight Ws of the seasonal variation and the weight Wt of the trend variation are added.

As an example, as shown in Equation 3 below, the reevaluation search score R (i) includes the search score S (i) of the i-th information of the input keyword search result, and the search result seasonal keyword included therein As a weight for each of the seasonal variation Ms (j) and the trend variation Mt (j) with respect to (j), the sum is obtained by multiplying the seasonal variation weight Ws and the trend variation weight Wt.

Further, the method of calculating the re-evaluation search score R (i) according to Equation 3 is an example, and depending on the setting of the search score S (i), as shown in Equation 4, the search score S (i) and the seasonal variation amount Ms. You may obtain | require by multiplying (j), trend fluctuation amount Mt (j), etc.

  The search result seasonal keyword (j) is only for those that match or are similar to the input keyword, and the re-evaluation search score is adjusted by adding the weight Ws of seasonal variation and the weight Wt of trend variation. R (i) may be calculated, or the seasonal variation weight Ws and the trend variation weight Wt may be added to general seasonal keywords that are not limited thereto. That is, the search result seasonal keyword obtained by the re-evaluation unit 204 may target all the seasonal keywords stored in the seasonal keyword information storage unit, as described above, or input keywords You may limit to the seasonal keyword relevant to the input keyword to include. It should be noted that the relationship between the search result seasonal keyword and the input keyword is based on a method using a synonym dictionary (thesaurus), such as “Nabe” and “Yosenabe”, which have a high percentage of matching characters. This can be determined by a general method.

  Here, the seasonal variation weight Ws and the trend variation weight Wt will be described with reference to FIG. FIG. 5 shows a specific example of the keyword seasonal variation weighting Ws and the trend variation weighting Wt.

  User requests such as whether the user wants to obtain information that emphasizes seasonality, information that emphasizes trendiness, information that does not place importance on seasonality, or information that does not place importance on trendiness Thus, the weighting Ws of seasonal variation and the weighting Wt of trend variation are set to predetermined values using a GUI or the like. That is, if the weight Ws of the seasonal variation and the weight Wt of the trend variation are set to be larger than the weight of the variation that the user does not want to emphasize the variation weight, the re-evaluation search score The user's request is reflected in R (i). Further, if the keyword fluctuation amount that is not to be emphasized is set so that the weighting is a negative value, the fluctuation amount can be added to the negative value, so that the user's request can be effectively reflected.

FIG. 5 is an example showing a user's request for a keyword related to a pan. A user's request | requirement exists in the information regarding a pan, excluding the information regarding seasonality, and searching for the information buried in seasonality. In this case, as shown in FIG. 5, it may be determined so as to eliminate the weight Ws of the seasonal variation and increase the weight Wt of the trend variation. In addition, regarding “Yosenabe” and “Fugu Nabe” that feel the season, the weighting of information including seasonality factors is reduced by setting the weighting Ws of seasonal variation to a negative value.

Thus, by determining the weighting Ws of the seasonal fluctuation amount and the weighting Wt of the trend fluctuation amount, the information requested by the user can be obtained quickly and easily. Note that the seasonal variation weight Ws and the trend variation weight Wt are set for each of a plurality of keywords related to the pan, but may be set only for the pan (input keyword), or the seasonal variation weight Ws or It may be set to only one of the trend variation amounts weighting Wt.

FIG. 6 shows the re-evaluation unit 204, from the search score S (i) of the search result shown in FIG. 4, the seasonal keyword shown in FIG. 3, its seasonal fluctuation amount Ms and the trend fluctuation amount Mt, and the keyword shown in FIG. The reevaluation search score R (i), which is an evaluation index, is calculated by substituting the seasonal variation amount weighting Ws and the trend variation weighting Wt into Equation 3, and the reevaluation search score R (i) is large. The results are shown in order. That is, the re-evaluation unit 204 obtains the result of the input keyword search (input keyword search result) by the general search engine 206 of FIG. 4 as the seasonal variation weight Ws and the trend variation weight Wt for the seasonal keyword. The search result is re-evaluated in consideration of the importance of the seasonality factor or the trendiness factor desired by the user shown in (1).

For example, in the case of the mushi pot, if the search result seasonal keyword j = 1 is “mushi pot” and j = 2 is “nabe”, the seasonal variation weighting Ws and trend for “mushi pot” shown in FIG. Multiplying the weight Wt of the fluctuation amount by the seasonal fluctuation amount Ms and the trend fluctuation amount Mt related to the “rice pan” shown in FIG. 3, the weighting Ws of the seasonal fluctuation amount and the weighting of the trend fluctuation amount related to the “hot pot” shown in FIG. The re-evaluation search score R (i) is calculated by multiplying Wt by the seasonal variation amount Ms and the trend variation amount Mt related to “nabe” shown in FIG. As for the pan lid, the seasonal variation amount Ms, the trend variation amount Mt and the seasonal variation regarding the “pan lid” and “pan” shown in FIGS. The reevaluation search score R (i) is calculated by multiplying the weight Ws of the quantity and the weight Wt of the trend variation. Similarly, the search result seasonal keyword j of Fugu Nabe is “Fugu Nabe” and “Nabe”, and the search result seasonal keyword j of Yosenabe is “Yose Nabe” and “Nabe”, and the re-evaluation search score R (i) Calculate and re-evaluate.

  As shown in FIG. 6, information related to “steaming pot” and “pan lid” that may have been overlooked and missed in an input keyword search by a general search engine 206 is searched for in the upper part. Will come to be. Further, in the search result by the general search engine 206, information such as “Yosenabe” and “Fugu nabe” searched for in the upper rank is searched in the lower rank according to the weighting requested by the user. Information that does not depend on the required seasonality can be obtained quickly.

  As described above, by using the search device according to the embodiment of the present invention, information that is buried in a keyword search by the conventional general search engine 206 can be shown at the top, so that information useful to the user can be easily obtained. Can be searched. The re-evaluation search scores R (i) are rearranged in descending order, but may be arranged in ascending order according to the user's request.

  In the method of calculating the re-evaluation search score R (i) using the formulas 3 and 4, if both the seasonal variation weight Ws and the trend variation weight Wt are set to 0, an existing search engine, etc. The search score S (i) and the reevaluation search score R (i) have the same value. Therefore, it is also possible to display a search result by a conventional search engine or the like before reflecting the user's request as one result of the reevaluation search score R (i).

  In the present embodiment, keywords related to pans and pans are reevaluated as search result seasonal keywords based on the input keyword search results by a general search engine 206 with “nabe” as an input keyword. Naturally, it is also possible to re-evaluate the result of the input keyword search by the search engine 206 using the sex keyword (determination of the search result seasonal keyword and calculation of the reevaluation search score R). In a search using a general seasonal keyword, a plurality of seasonal keywords included in the input keyword search result are used as search result seasonal keywords, and the search is performed again with the seasonal variation weight Ws and the trend variation weight Wt added. It is also possible to obtain information that is desired.

  As described above, a search that takes into account the seasonality included in the keyword or excludes the seasonality can be performed based on the search by the keyword. Information that has been ranked and missed is ranked higher, and keyword search can be performed efficiently.

(5) PC Hardware Configuration Here, the PC hardware configuration will be described with reference to FIG.

  The PC 700 includes a ROM 705, a RAM 710, a CPU 715, a bus 720, an internal interface (internal I / F) 725, and an external interface (external I / F) 730. The ROM 705 stores a basic program executed by the PC 700, a program that starts up in the event of an abnormality, and the like. The RAM 710 stores various programs such as reevaluation processing and result display processing, and data such as keyword information. The ROM 705 and the RAM 710 are examples of a storage device, and may be a storage device such as an EEPROM, an optical disk, or a magneto-optical disk.

  The CPU 715 executes search result re-evaluation processing and result display processing according to various programs such as a re-evaluation processing program. A bus 720 is a path for exchanging data among the devices such as the ROM 705, the RAM 710, the CPU 715, the internal interface 725, and the external interface 730.

  The internal interface 725 inputs data such as keywords, and outputs necessary data to a display and a speaker (not shown). The external interface 730 transmits and receives data to and from connected devices such as a server and other PCs. For example, the search result after the re-evaluation may be displayed on a display provided on the PC 700 via the internal interface 725 or a display provided on another PC connected to the network via the external interface 730. May be displayed.

(6) Flow of Display Processing According to One Embodiment FIG. 8 is a flowchart showing the flow of keyword search executed in the display processing according to one embodiment of the present invention.

As shown in FIG. 8, when the search process is started (S800), an input keyword search is performed by a general search engine 206 by a user input, and the input keyword search result is stored in the search result storage unit 203 (S801). Subsequently, it is determined whether to display the input keyword search result in S801 (S802). When re-evaluating the search result by the input keyword search so as to meet the user's request, the search result re-evaluation process of the seasonal keyword search is performed without confirming the result by the general input keyword search (S803). ). The re-evaluation in S803 is performed by the re-evaluation unit 204, and as described in detail as a re-evaluation process described later, the seasonal keyword is obtained from the search score S included in the input keyword search result stored in the search result storage unit 203. The re-evaluation search score R (i) is weighted so that the user's request is added to the seasonality and trend of the seasonal keyword or the keyword related to the input keyword among the seasonal keywords stored in the information storage unit 202. ) Is calculated. Next, sorting is performed again based on the value of the reevaluation search score R (i). Subsequently, the re-evaluated result is displayed on the display unit 205 (S804). As a result, if the information requested by the user can be obtained, this processing is terminated without further re-evaluation (S808). On the other hand, if you want to re-evaluate the results of seasonal keyword search from a different point of view again, repeat the re-evaluation until the information requested by the user is obtained by changing the weighting, which is the importance for seasonal factors and trend factors. This process is terminated when no re-evaluation is requested (S808).

  In addition, when the user requests display of the search result by the input keyword search by the general search engine 206 (S802), the result is displayed (S806), and the user is made to judge whether re-evaluation is possible or not based on the result ( (S807) If the re-evaluation is requested, the processes of S803 to S805 are repeated, and if not requested, the process is terminated (S808).

(7) Flow of Re-evaluation Processing According to One Embodiment FIG. 9 is a flowchart showing the flow of re-evaluation executed in the re-evaluation processing according to one embodiment of the present invention. The flow of the reevaluation unit 204 will be described with reference to FIG.

  As shown in FIG. 9, the re-evaluation unit 204 weights the seasonal variation amount Ws and the trend variation amount weight Wt, which are the importance of the seasonality factor and the trendiness factor that the user requests for the search result seasonality keyword. (S901). Subsequently, the seasonal variation weight Ws and the trend variation weight Wt captured in S901 and the seasonal variation amount Ms and trend variation amount Mt of each seasonal keyword stored in the seasonal keyword information storage unit 202 are used. The re-evaluation search score R (i) is calculated from the search score S included in the search result stored in the search result storage unit 203 (S902). Subsequently, the search articles are sorted in descending order of the value of the reevaluation search score R (i) calculated in S902 (S903).

  As described above, according to the search device according to the present embodiment, information required by the user can be efficiently searched, and information useful for the user can be easily searched.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, this invention is not limited to the example which concerns. It is obvious for those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea described in the claims, and of course, the technical scope of the present invention is also possible. It is understood that it belongs to.

It is a figure explaining time series data. It is explanatory drawing which shows the function structure of the search device which concerns on one Embodiment of this invention. It is a figure which shows the specific example accumulate | stored in the seasonal keyword information storage part which concerns on one Embodiment of this invention. It is a figure which shows the result of the input keyword search by a general search engine. It is a figure which shows an example of the weight Ws of the seasonal variation | change_quantity based on one Embodiment of this invention, and the weight of a trend variation. It is a figure which shows the result of the reevaluation which concerns on one Embodiment of this invention. It is a figure which shows the hardware constitutions of PC. It is a flowchart which shows the flow of the search device performed in the search device which concerns on one Embodiment of this invention. It is a flowchart which shows the flow performed in the reevaluation process which concerns on one Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 200 Search apparatus 201 Fluctuation extraction part 202 Seasonal keyword information storage part 203 Search result storage part 204 Re-evaluation part 205 Display part 206 Search engine 700 PC
705 ROM
710 RAM
715 CPU
720 bus 725 internal interface 730 external interface

Claims (9)

Analyzing keyword time-series data to extract seasonal variation and trend variation,
A seasonal keyword information storage unit which accumulates the seasonal variation amount and trend variation amount extracted by the variation extraction unit in association with keywords;
A search result storage unit for storing an input keyword search result based on at least one input keyword set by a user;
The result of the input keyword search stored in the search result storage unit is again stored in the seasonal keyword information storage unit using a weight for at least one of the seasonal variation amount and the trend variation amount set by the user. An input keyword search stored in the search result storage unit based on the weighted seasonal keyword search result by performing a search for the seasonal keyword weighted to the seasonal variation amount or trend variation amount of the accumulated seasonal keyword A re-evaluation unit that sorts the results of
A search device comprising: The seasonal keyword search weighted by the seasonal variation amount or trend variation amount performed by the re-evaluation unit is stored in the seasonal keyword information storage unit, and the seasonal keyword related to the input keyword including the input keyword is used. The search device according to claim 1. The search result stored in the search result storage unit includes a search score,
The re-evaluation unit calculates the seasonal variation amount and trend variation amount of the seasonal keyword accumulated in the seasonal keyword information storage unit, the weight of seasonal variation amount set by the user, and the trend variation amount from the search score. The search device according to claim 1, wherein a re-evaluation search score is calculated using weighting.   The search device according to claim 1, wherein the weighting of the seasonal variation amount and the trend variation amount set by the user includes a negative weighting.   The search device according to claim 3 or 4, wherein the search score is a value obtained by searching for an input keyword, a reciprocal value corresponding to the search order, or a value obtained by switching the high order and low order of the search order.   The search device according to claim 1, further comprising a display unit that displays a search result by the re-evaluation unit. A search engine for performing an input keyword search using at least one input keyword set by a user;
A search result storage unit for storing a result of an input keyword search by the search engine;
Analyzing keyword time-series data to extract seasonal variation and trend variation,
A seasonal keyword information storage unit which accumulates the seasonal variation amount and trend variation amount extracted by the variation extraction unit in association with keywords;
The result of the input keyword search stored in the search result storage unit is again stored in the seasonal keyword information storage unit using a weight for at least one of the seasonal variation amount and the trend variation amount set by the user. Search for a seasonal keyword weighted to the seasonal variation amount or trend variation amount of the accumulated seasonal keyword, and input keyword search stored in the search result storage unit based on the weighted seasonal keyword search result A re-evaluation unit that sorts the results of
A search system comprising: Analyzing the keyword time-series data to extract seasonal variation and trend variation,
Accumulating seasonal keyword information extracted by the extraction step;
A search step for performing an information search using input keywords and capturing the search results;
Capturing a weight for at least one of seasonal variation and trend variation set by the user;
Calculating a re-evaluation search score using the weighting set by the user in the search result and the seasonal keyword information;
Reordering the reevaluation search score based on the value of the reevaluation search score;
A search method comprising:   The search method according to claim 8, further comprising a display step of displaying a search result obtained by the re-evaluation step.

Images