Customer attraction support device, customer attraction support system, customer attraction support method, and program

Abstract

The present invention supports customer attraction by quantifying the strength of keywords, topics, or the like that have been input, predicting the number of attracted customers, the number of clicks, or the like, and selecting keywords suitable for a website or a guide text, for example. The present invention provides a customer attraction support device comprising an input processing unit 11a that accepts the input of a keyword and processes the keyword, a strength determination unit 11d that determines the strength of the keyword, a trend correction unit 11e that performs trend correction of the determination result from the strength determination unit, a prediction unit 11f that predicts at least the number of attracted customers on the basis of the determination result after the trend correction, and an output unit 11h that outputs the prediction result, wherein the prediction unit 11f performs the prediction using a mathematical model obtained by machine learning that employs at least one of seminar information, mail information, and site information as training data.

Description

Customer attraction support device, customer attraction support system, customer attraction support method, and program

[0001] The present invention relates to a technology for quantifying the strength of input keywords, issues, etc., predicting the number of customers and clicks, etc., and supporting customer acquisition.

[0002] Conventionally, the keywords used to compose the content of a website or the like have a significant impact on the number of visitors and clicks to the website, so the keywords have been selected based on the results of various marketing surveys.

[0003] On the other hand, when planning seminars and events, it is also desirable to predict the number of attendees that can be expected at the time of planning and to provide effective information that will attract attendees.

[0004] For example, Patent Document 1 discloses an order prediction device that includes an acquisition means for acquiring event data related to an event, a prediction means for predicting the attention level and order probability for each data item included in the event data using an attention mechanism based on the event data, and an output means for outputting the attention level and order probability as prediction results.

[0005] International Publication No. WO2022 / 201512A1

[0006] However, conventional technologies such as Patent Document 1 do not assist in attracting customers by quantifying the strength of input keywords, issues, etc., predicting the number of customers and clicks, etc., and selecting keywords suitable for websites, informational text, etc.

[0007] The present invention was made in consideration of these problems, and its purpose is to provide a technology that supports attracting customers by quantifying the strength of input keywords, issues, etc., predicting the number of customers and clicks, etc., and selecting keywords that are suitable for websites, informational text, etc.

[0008] In order to solve the above-mentioned problems, a customer attraction support device according to a first aspect of the present invention includes an input processing unit that receives and processes input of keywords, a strength determination unit that determines the strength of the keywords, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a prediction unit that predicts at least the number of customers to be attracted based on the determination result after the trend correction, and an output unit that outputs the prediction result, wherein the strength determination unit processes text data including seminar information related to a seminar announcement having a title that has a high similarity to the text data processed by the input processing unit, the seminar information including a title, a lead sentence, the number of applicants, and the number of attendees, and a mail newsletter including a title, a body, an open rate, and a click rate. A predetermined number of email information including the email address and the email message, and site information including the title, body, PV value, and viewing time related to a website are extracted, and a weighting coefficient related to type and a weighting coefficient related to time are set for each of them, and a numerical strength is calculated by calculating an average value using the weighting coefficients.The prediction unit makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of seminar information including the title, lead text, number of applicants, and number of attendees related to a seminar announcement, email information including the title, body, open rate, and click rate related to an email newsletter, and site information including the title, body, PV value, and viewing time related to a website.

[0009] A customer attraction support device according to a second aspect of the present invention comprises an input processing unit that accepts and processes keyword input, a strength determination unit that determines the strength of the keyword, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a generation unit that generates content based on the determination result after the trend correction, and an output unit that outputs the content.

[0010] A customer attraction support system according to a third aspect of the present invention is a customer attraction support system including a user's terminal device and a customer attraction support device, wherein the user's terminal device comprises a first transmission unit that transmits keywords to the customer attraction support device, a first reception unit that receives prediction results from the customer attraction support device, and a display unit that displays the prediction results, and the customer attraction support device comprises a second reception unit that receives keywords from the user's terminal device, an input processing unit that accepts and processes input of the keywords, a strength determination unit that determines the strength of the keywords, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a prediction unit that predicts at least the number of customers to be attracted based on the determination result after the trend correction, and a second transmission unit that transmits the prediction results to the user's terminal device, wherein the strength determination unit is configured to determine the strength of the keywords based on the determination result after the trend correction A predetermined number of pieces of seminar information including the title, lead text, number of applicants, and number of attendees related to seminar announcements having a title, email information including the title, body text, open rate, and click rate related to email newsletters, and site information including the title, body text, PV value, and viewing time related to websites are extracted, and a weighting coefficient related to type and a weighting coefficient related to time are set for each, and a numerical strength is calculated by calculating an average value using the weighting coefficients.The prediction unit makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of the seminar information including the title, lead text, number of applicants, and number of attendees related to seminar announcements, the email information including the title, body text, open rate, and click rate related to email newsletters, and site information including the title, body text, PV value, and viewing time related to websites.

[0011] A fourth aspect of the present invention provides a customer attraction support system comprising a user's terminal device and a customer attraction support device, wherein the user's terminal device comprises a first transmitting unit that transmits keywords to the customer attraction support device, a first receiving unit that receives content from the customer attraction support device, and a display unit that displays the content, and the customer attraction support device comprises a second receiving unit that receives keywords from the user's terminal device, an input processing unit that accepts and processes input of the keywords, a strength determination unit that determines the strength of the keywords, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a generation unit that generates content based on the determination result after the trend correction, and a second transmitting unit that transmits the content to the user's terminal device.

[0012] A customer attraction support method according to a fifth aspect of the present invention includes an input processing unit that receives and processes an input of a keyword, a strength determination unit that determines the strength of the keyword, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a prediction unit that predicts at least the number of customers to be attracted based on the determination result after the trend correction, and an output unit that outputs the prediction result, wherein the strength determination unit outputs seminar information including a title, a lead sentence, the number of applicants, and the number of attendees related to a seminar announcement having a title that has a high similarity to the text data processed by the input processing unit, email information including a title, a body, an open rate, and a click rate related to an email newsletter, and A predetermined number of pieces of site information related to the b site, including the title, main text, PV value, and viewing time, are extracted, and a weighting coefficient related to type and a weighting coefficient related to time are set for each, and a numerical strength is calculated by calculating an average value using the weighting coefficients.The prediction unit makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of seminar information related to seminar announcements, including the title, lead text, number of applicants, and number of attendees, email information related to email newsletters, including the title, main text, open rate, and click rate, and site information related to websites, including the title, main text, PV value, and viewing time.

[0013] A customer attraction support method according to a sixth aspect of the present invention includes an input processing unit that accepts and processes keyword input, a strength determination unit that determines the strength of the keyword, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a generation unit that generates content based on the determination result after the trend correction, and an output unit that outputs the content.

[0014] A program according to a seventh aspect of the present invention causes a computer to function as an input processing unit that receives and processes input of keywords, a strength determination unit that determines the strength of the keywords, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a prediction unit that predicts at least the number of attendees based on the determination result after the trend correction, and an output unit that outputs the prediction result, wherein the strength determination unit performs the following: A predetermined number of pieces of email information and site information relating to websites, including titles, text, PV values, and viewing times, are extracted, and a weighting coefficient relating to type and a weighting coefficient relating to time are set for each, and a numerical strength is calculated by calculating an average value using the weighting coefficients.The prediction unit makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of seminar information relating to seminar announcements, including titles, lead text, number of applicants, and number of attendees, email information relating to email newsletters, including titles, text, open rates, and click rates, and site information relating to websites, including titles, text, PV values, and viewing times.

[0015] A program according to an eighth aspect of the present invention causes a computer to function as an input processing unit that accepts and processes keyword input, a strength determination unit that determines the strength of the keyword, a trend correction unit that performs trend correction on the determination result of the strength determination unit, a generation unit that generates content based on the determination result after the trend correction, and an output unit that outputs the content.

[0016] According to the present invention, it is possible to provide a technology that assists in attracting customers by quantifying the strength of input keywords, issues, etc., predicting the number of customers and clicks, etc., and selecting keywords that are suitable for websites, information text, etc.

[0017] FIG. 1 is a configuration diagram of a customer attraction support system according to an embodiment of the present invention. FIG. 2 is a configuration diagram of a customer attraction support device of the system. FIG. 3 is a configuration diagram of a terminal device of the system. FIG. 4 is a flowchart showing the procedure for prediction processing by the system. FIG. 5 is a flowchart showing the procedure for content generation by the system. FIGS. 6(a) to 6(d) are configuration examples of various tables managed by the customer attraction support device of the system. FIG. 7 is an explanatory diagram of the construction of a numerical model by the system. FIG. 8 is an explanatory diagram of trend correction by the system.

[0018] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0019] FIG. 1 shows the configuration of a customer attraction support system according to an embodiment of the present invention.

[0020] As shown in the figure, the customer attraction support system according to the embodiment of the present invention is configured by connecting a customer attraction support device 1, which is made up of a server device or the like, to a user's terminal device 2 via a communication network such as the Internet, either wirelessly or via a wired connection. The user's terminal device 2 can be a variety of devices, such as a smartphone, tablet, laptop computer, or desktop personal computer.

[0021] In this configuration, the customer attraction support device 1 accepts and processes text data such as keywords and assignments sent from the user's terminal device 2, determines the strength of the keywords, etc., performs trend correction on the determination result, and predicts and outputs at least the number of customers to be attracted based on the determination result after trend correction. This output form includes, for example, output to an external output device such as a monitor or printer, as well as sending data to the user's terminal device 2.

[0022] The customer attraction support device 1 also receives and processes text data such as keywords and assignments sent from the user's terminal device 2, determines the strength of the keywords, etc., performs trend correction on the determination result, and generates and outputs content based on the trend-corrected determination result. Again, the form of this output includes output to an external output device such as a monitor or printer, as well as sending data to the user's terminal device 2.

[0023] The customer attraction support device 1 constructs and stores a mathematical model obtained by machine learning using as learning data at least one of the following: seminar information including the title, lead text, number of applicants, and number of attendees related to seminar announcements; email information including the title, body text, open rate, and click rate related to email newsletters; and site information including the title, body text, PV value, and viewing time related to websites; and the mathematical model is used to make predictions about customer attendance.

[0024] FIG. 2 shows the configuration of the customer attraction support device in the customer attraction support system, and will be described.

[0025] As shown in the figure, the customer attraction support device 1 includes a control unit 11, a communication I / F 12, and a storage unit 13. The control unit 11, the communication I / F 12, and the storage unit 13 are communicatively connected via a system bus. In addition to these components, the device may include input units such as a keyboard and a mouse, and a display unit such as an LCD display for displaying various information. The communication I / F 12 is a communication interface, implemented, for example, by a NIC, connected to a communication network 4 such as the Internet via a wired or wireless connection, and communicating with the terminal device 2, etc. The storage unit 13 is implemented, for example, by a semiconductor memory device such as RAM or flash memory, a HDD, or an optical disk device, and stores programs to be executed by the control unit 11 in advance. The storage unit 13 also includes a user information storage unit 14, a seminar information storage unit 15, an email information storage unit 16, a site information storage unit 17, and a learning storage unit 18.

[0026] The user information storage unit 14 stores attribute information such as the name, age, gender, address, telephone number, and email address of registered users, linked to the user ID. The seminar information storage unit 15 stores information such as the title, lead sentence, date of the seminar, number of applicants, number of attendees, and application speed of the seminar information, linked to the seminar ID, as shown in FIG. 6( a). The email information storage unit 16 stores information such as the title, body, distribution date, open rate, and click rate of the email newsletter, linked to the email ID, as shown in FIG. 6( b). The site information storage unit 17 stores information such as the title, body, view date, PV value, and view time of the website, linked to the site ID, as shown in FIG. 6( c). The learning storage unit 17 stores and updates mathematical models used in strength determination.

[0027] The control unit 11 is realized by a CPU, an MPU, etc. The control unit 11 is composed of an integrated circuit such as an ASIC, an FPGA, etc. The control unit 11 reads and executes programs stored in the storage unit 13, and thereby functions as an input processing unit 11a, a learning unit 11b, a management unit 11c, a strength determination unit 11d, a trend correction unit 11e, a prediction unit 11f, a generation unit 11g, an output unit 11h, a transmission / reception control unit 11i, and a main control unit 11j.

[0028] In this configuration, the input processing unit 11a accepts and processes input of text data related to keywords and tasks. The learning unit 11b uses the data stored in the storage unit 13 as learning data to construct a mathematical model for predicting customer numbers and updates it as appropriate.

[0029] 7, the learning unit 11b executes machine learning using as learning data the "seminar information" including the title, lead text, date, number of applicants, number of attendees, and application speed of seminar announcements stored in the seminar information storage unit 15, the "email information" including the title, text, distribution date, open rate, and click rate of e-mail newsletters stored in the e-mail information storage unit 15, and the "site information" including the title, text, view date, PV value, and view time of websites stored in the site information storage unit 17, thereby constructing a mathematical model for customer attraction prediction and storing the model in the learning storage unit 17. This numerical model is updated as appropriate through periodic learning by the learning unit 11b.

[0030] The management unit 11c manages the user information in the user information storage unit 14, the seminar information in the seminar information storage unit 15, the mail information in the mail information storage unit 16, and the site information in the site information storage unit 17, and updates them as appropriate.

[0031] The strength determination unit 11d determines the strength of the text data such as keywords and assignments processed by the input processing unit 11a. More specifically, as shown in Fig. 6(d), a predetermined number of pieces of seminar information, email information, and site information having titles highly similar to the input processed text data are extracted, and a weighting coefficient related to type and a weighting coefficient related to time are set for each piece, and a numerical strength is calculated by calculating an average value using these weighting coefficients.

[0032] The trend correction unit 11e performs trend correction on the quantified strength calculated by the strength determination unit 11d. More specifically, as shown in Fig. 8, the trend correction unit 11e estimates the trend of the data from the process of time-series changes in the seminar information, email information, and site information, and corrects the value at the time of strength measurement to the value at the timing when the strength is actually desired to be calculated.

[0033] The prediction unit 11f uses the mathematical model constructed by the learning unit 11b to predict customer attraction marketing trends, such as the number of customers and the number of clicks, based on the strength after trend correction. The generation unit 11g narrows down the strongest keywords from the input keywords to the top few words based on the strength after trend correction, and generates content such as announcements using the keywords using generative AI. The output unit 11h outputs the strength indicators and content. The transmission / reception control unit 11i functions as a transmission unit (second transmission unit) that transmits data related to the strength indicators and content in HTML format to the user's terminal device 2, and a reception unit (second reception unit) that receives text data such as keywords and assignments from the user's terminal device 2. In addition, the main control unit 11j is responsible for other control of the customer attraction support device 1.

[0034] FIG. 3 shows the configuration of the terminal device in the customer attraction support system.

[0035] As shown in the figure, the terminal device 2 has a control unit 21, a communication I / F 22, an operation unit 23, a display unit 24, and a memory unit 25. Each unit 21 to 25 is connected to each other via a system bus for communication. The communication I / F 22 is implemented, for example, by a NIC or the like, and is connected to a communication network 3 such as the Internet via a wired or wireless connection, and serves as a communication interface for communicating with the customer attraction support device 1 and the like. The operation unit 23 is implemented, for example, by a mouse or keyboard, and accepts various operation inputs from the user. The display unit 24 is implemented, for example, by a liquid crystal display, and displays various information. The operation unit 23 and the display unit 24 may be integrated into a touch panel. The memory unit 25 is implemented, for example, by a semiconductor memory device such as RAM or flash memory, a HDD, or an optical disk device, and stores the program executed by the control unit 21.

[0036] The control unit 21 is realized by a CPU, an MPU, or the like, and may be configured as an integrated circuit such as an ASIC or an FPGA. The control unit 21 functions as a transmission / reception control unit 21a and a display control unit 21b by reading and executing programs stored in the storage unit 25. The transmission / reception control unit 21a functions as a transmission unit (first transmission unit) that transmits text data such as keywords and assignments input by operating the operation unit 23 to the customer attraction support device 1, and a reception unit (first reception unit) that receives data in HTML format or the like sent from the customer attraction support device 1. The display control unit 21b controls the display unit 24 to display a predetermined content based on the data in HTML format or the like received by the transmission / reception control unit 21a.

[0037] The procedure of the prediction process performed by the customer attraction support device will be described in detail below with reference to the flowchart of FIG.

[0038] When processing begins, the input processing unit 11a accepts and processes input of text data related to keywords and assignments (S1). Subsequently, the strength determination unit 11d determines the strength of the text data, such as the keywords and assignments, processed by the input processing unit 11a (S2). More specifically, as shown in Fig. 6(d), a predetermined number of seminar information, email information, and website information having titles highly similar to the input text data are extracted, and a weighting factor related to type and a weighting factor related to time are assigned to each of them, and a numerical strength is calculated by calculating an average value using these weighting factors.

[0039] Next, the trend correction unit 11e performs trend correction on the quantified strength calculated by the strength determination unit 11d (S3). More specifically, as shown in Fig. 8, for example, the trend of the data is estimated from the process of time-series changes in the seminar information, email information, and site information, and the value at the time of strength measurement is corrected to the value at the timing when the strength is actually desired to be calculated.

[0040] Next, the prediction unit 11f predicts customer attraction marketing trends such as the number of customers attracted and the number of clicks based on the strength after trend correction using the mathematical model constructed by the learning unit 11b (S4), and the output unit 11h outputs an index of strength (S5). In this way, the series of processes related to customer attraction prediction is completed.

[0041] Next, the content generation procedure by the customer attraction system will be described in detail with reference to the flowchart of FIG.

[0042] When the process starts, the input processing unit 11a accepts and processes input of text data related to keywords and assignments (S11). Subsequently, the strength determination unit 11d determines the strength of the text data, such as the keywords and assignments, processed by the input processing unit 11a (S12). More specifically, as shown in FIG. 6(d), a predetermined number of pieces of seminar information, email information, and website information having titles highly similar to the input text data are extracted, and weighting factors related to type and time are assigned to each piece. The weighting factors are used to calculate an average value, thereby calculating a numerical strength.

[0043] Next, the trend correction unit 11e performs trend correction on the quantified strength calculated by the strength determination unit 11d (S13). More specifically, as shown in FIG. 8, for example, the trend of the data is estimated from the time-series changes in the seminar information, email information, and site information, and the value at the time of strength measurement is corrected to the value at the timing when the strength is actually desired to be calculated. Then, the generation unit 11g narrows down the strongest keywords from the multiple input keywords to the top few words based on the strength after trend correction, and generates content such as an announcement message using the keywords using a generative AI (S15), and the output unit 11h outputs the content (S15).

[0044] As described above, according to the embodiment of the present invention, the strength of an input keyword in terms of attracting customers can be quantified and calculated, and based on that strength, a mathematical model obtained by learning can be used to predict marketing trends related to attracting customers (e.g., the number of customers attracted, the number of clicks, etc.), thereby providing indirect support for attracting customers. Furthermore, based on the calculated strength, keywords that are effective in attracting customers can be selected, and information such as seminar invitations using those keywords can be automatically generated by generative AI, thereby providing indirect support for attracting customers in this respect as well.

[0045] Although the embodiments of the present invention have been described above, it goes without saying that the present invention is not limited to these and that various improvements and modifications can be made without departing from the spirit and scope of the present invention.

[0046] For example, the learning data used to build a mathematical model for predicting customer numbers is not limited to the seminar information, email information, and website information mentioned above, but various data can be added as learning data depending on the purpose, and a numerical model can be built.

[0047] 1...customer attraction support device, 2...terminal device, 3...communication network, 11...control unit, 11a...input processing unit, 11b...learning unit, 11c...management unit, 11d...strength determination unit, 11e...trend correction unit, 11f...prediction unit, 11g...generation unit, 11h...output unit, 11i...transmission and reception control unit, 11j...main control unit, 12...communication I / F, 13...memory unit, 14...user information memory unit, 15...seminar information memory unit, 16...email information memory unit, 17...site information memory unit, 18...learning memory unit, 21...control unit, 21a...transmission and reception control unit, 21b...display control unit, 22...communication I / F, 23...operation unit, 24...display unit, 25...memory unit.

Claims

1. A system comprising: an input processing unit that receives and processes keyword input; a strength judgment unit that judges the strength of the keyword; a trend correction unit that performs trend correction on the judgment result of the strength judgment unit; a prediction unit that predicts at least the number of attendees based on the judgment result after the trend correction; and an output unit that outputs the prediction result, wherein the strength judgment unit extracts a predetermined number of pieces of seminar information including a title, lead text, number of applicants, and number of attendees related to seminar invitations having titles that have a high similarity to the text data processed by the input processing unit, email information including a title, body text, open rate, and click rate related to email newsletters, and site information including a title, body text, PV value, and viewing time related to websites, sets a weighting coefficient related to type and a weighting coefficient related to time for each, and calculates a numerical strength by calculating an average value using the weighting coefficients, The prediction unit makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of seminar information related to seminar invitations, including the title, lead text, number of applicants, and number of attendees, email information related to email newsletters, including the title, body text, open rate, and click rate, and site information related to websites, including the title, body text, PV value, and viewing time.

2. A customer attraction support device comprising: an input processing unit that accepts and processes keyword input; a strength judgment unit that judges the strength of the keywords; a trend correction unit that performs trend correction on the judgment result of the strength judgment unit; a generation unit that selects keywords effective for attracting customers based on the judgment result after the trend correction and generates content; and an output unit that outputs the content, wherein the strength judgment unit extracts a predetermined number of pieces of seminar information including a title, lead text, number of applicants, and number of attendees related to seminar invitations, which have titles that are highly similar to the text data processed by the input processing unit; email information including a title, body text, open rate, and click rate related to email newsletters; and site information including a title, body text, PV value, and viewing time related to websites, and sets a weighting coefficient related to type and a weighting coefficient related to time for each of the pieces, and calculates a numerical strength by calculating an average value using the weighting coefficients.

3. A customer attraction support system comprising a user's terminal device and a customer attraction support device, wherein the user's terminal device comprises: a first transmission unit which transmits keywords to the customer attraction support device; a first reception unit which receives prediction results from the customer attraction support device; and a display unit which displays the prediction results; and the customer attraction support device comprises: a second reception unit which receives keywords from the user's terminal device; an input processing unit which accepts and processes input of the keywords; a strength determination unit which determines the strength of the keywords; a trend correction unit which performs trend correction on the determination result of the strength determination unit; a prediction unit which predicts at least the number of customers to be attracted based on the determination result after the trend correction; and a second transmission unit which transmits the prediction result to the user's terminal device, The strength determination unit extracts a predetermined number of pieces of seminar information including a title, lead text, number of applicants, and number of attendees related to seminar invitations, which have titles highly similar to the text data processed by the input processing unit, email information including a title, body text, open rate, and click rate related to e-mail magazines, and site information including a title, body text, PV value, and viewing time related to websites, sets a weighting coefficient related to type and a weighting coefficient related to time for each of the pieces, and calculates a numerical strength by calculating an average value using the weighting coefficients.The prediction unit makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of the seminar information including the title, lead text, number of applicants, and number of attendees related to seminar invitations, the email information including the title, body text, open rate, and click rate related to e-mail magazines, and site information including the title, body text, PV value, and viewing time related to websites.

4. A customer attraction support system comprising a user's terminal device and a customer attraction support device, wherein the user's terminal device comprises a first transmission unit which transmits keywords to the customer attraction support device, a first reception unit which receives content from the customer attraction support device, and a display unit which displays the content, and the customer attraction support device comprises a second reception unit which receives keywords from the user's terminal device, an input processing unit which accepts and processes input of the keywords, a strength determination unit which determines the strength of the keywords, a trend correction unit which performs trend correction on the determination result of the strength determination unit, a generation unit which selects keywords effective for attracting customers and generates content based on the determination result after the trend correction, and a second transmission unit which transmits the content to the user's terminal device, The strength determination unit extracts a predetermined number of pieces of seminar information including a title, lead text, number of applicants, and number of attendees related to seminar invitations, which have titles that are highly similar to the text data processed by the input processing unit, email information including a title, body text, open rate, and click rate related to email newsletters, and site information including a title, body text, PV value, and viewing time related to websites, sets a weighting coefficient related to type and a weighting coefficient related to time for each, and calculates a numerical strength by calculating an average value using the weighting coefficients.

5. An input processing unit receives and processes keyword input; a strength determination unit determines the strength of the keyword; a trend correction unit performs trend correction on the determination result of the strength determination unit; a prediction unit predicts at least the number of attendees based on the determination result after the trend correction; an output unit outputs the prediction result; and the strength determination unit extracts a predetermined number of pieces of seminar information including a title, lead text, number of applicants, and number of attendees related to seminar invitations having titles that are highly similar to the text data processed by the input processing unit, email information including a title, body text, open rate, and click rate related to email newsletters, and site information including a title, body text, PV value, and viewing time related to websites, sets a weighting coefficient related to type and a weighting coefficient related to time for each, and calculates a numerical strength by calculating an average value using the weighting coefficients, The prediction unit makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of seminar information related to seminar invitations, including the title, lead text, number of applicants, and number of attendees, email information related to email newsletters, including the title, text, open rate, and click rate, and site information related to websites, including the title, text, PV value, and viewing time.

6. A method for supporting customer attraction, in which an input processing unit accepts and processes keyword input, the strength determination unit extracts a predetermined number of pieces of seminar information relating to seminar invitations, the seminar information including titles, lead text, number of applicants, and number of attendees, email information relating to email newsletters, the email information including titles, body text, open rates, and click rates, and site information relating to websites, the email information including titles, body text, PV values, and viewing times, and sets a weighting coefficient related to type and a weighting coefficient related to time for each piece, and calculates a numerical strength by calculating an average value using the weighting coefficients, the trend correction unit performs trend correction on the judgment result of the strength judgment unit, the generation unit selects keywords that are effective in attracting customers based on the judgment result after trend correction and generates content, and the output unit outputs the content.

7. A computer is caused to function as an input processing unit that receives and processes keyword input, a strength judgment unit that judges the strength of the keyword, a trend correction unit that performs trend correction on the judgment result of the strength judgment unit, a prediction unit that predicts at least the number of attendees based on the judgment result after the trend correction, and an output unit that outputs the prediction result, wherein the strength judgment unit extracts a predetermined number of pieces of seminar information including the title, lead text, number of applicants, and number of attendees related to seminar invitations having titles that have a high similarity to the text data processed by the input processing unit, email information including the title, body text, open rate, and click rate related to email newsletters, and site information including the title, body text, PV value, and viewing time related to websites, sets a weighting coefficient related to type and a weighting coefficient related to time for each, and calculates a numerical strength by calculating an average value using the weighting coefficients, The prediction unit is a program that makes the above prediction using a mathematical model obtained by machine learning using as learning data at least one of seminar information including the title, lead text, number of applicants, and number of attendees related to seminar announcements, email information including the title, body text, open rate, and click rate related to email newsletters, and site information including the title, body text, PV value, and viewing time related to websites.

8. A program that causes a computer to function as: an input processing unit that accepts and processes keyword input; a strength determination unit that extracts a predetermined number of pieces of seminar information including the title, lead text, number of applicants, and number of attendees related to seminar invitations, which have titles that are highly similar to the text data processed by the input processing unit; email information including the title, body, open rate, and click rate related to email newsletters; and site information including the title, body, PV value, and viewing time related to websites, and sets a weighting coefficient related to type and a weighting coefficient related to time for each of the pieces and calculates a numerical strength by calculating an average value using the weighting coefficients; a trend correction unit that performs trend correction on the determination result of the strength determination unit; a generation unit that selects keywords that are effective in attracting customers based on the determination result after the trend correction and generates content; and an output unit that outputs the content.

Images