JP2019164408A - Recommendation system and program - Google Patents

Abstract

To provide a recommendation system and a program for providing information which provides a novel experience for a user.SOLUTION: A recommendation system comprises: information acquisition means 20 which acquires information about a user; character analyzing means 21 which derives character information on the user from the information acquired by the information acquisition means 20; and activity information extraction means 22 which extracts and outputs activity information, from activity information storage means 33 in which at least the activity information and the character information relevant to the activity information are stored in association with each other, based on the character information on the user derived by the character analyzing means 21. The activity information extraction means 22 performs scoring on the basis of compatibility between the character information relevant to the activity information and the character information on the user derived by the character analyzing means 21, about each activity information stored in the activity information storage means 33, and extracts and outputs both of the activity information with a high score and the activity information with a low score.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a recommendation system and a program that provide recommended information according to the personality of a user.

  In shopping sites, travel sites, and the like, recommendation systems that provide recommended information according to user preferences are widely used. In the recommendation system, a user's preference is estimated based on information about the user such as a purchase history and a browsing history, and information matching the user's preference is extracted and provided as recommended information.

  When extracting recommended information based on information about the user, first, information for scoring is derived from the information about the user, and the information stored in the database is scored from the information to obtain the score. Output high information as a result. As such a recommendation system, there exists a thing as listed in patent document 1, for example.

JP 2013-25325 A

  For example, when recommending an activity at a travel destination, consider using the above recommendation system. Activities at travel destinations may be selected depending on the user's usual hobbies and preferences, but users who want to experience an extraordinary feeling that can only be achieved by travel want activities that deviate from their usual hobbies and preferences Sometimes. In such a case, there may be a case where sufficient satisfaction cannot be obtained simply by providing recommended information that is highly compatible with information related to the user.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a recommendation system and a program that provide recommended information that brings a new experience to the user.

  A recommendation system according to the present invention that achieves the above object includes an information acquisition unit that acquires information about a user, a personality analysis unit that derives personality information of the user from the information acquired by the information acquisition unit, and the personality analysis unit. Activity information extracting means for extracting and outputting the activity information from activity information storage means stored in association with at least activity information and personality information related to the activity information based on the derived personality information of the user; The activity information extraction means includes, for each of the activity information stored in the activity information storage means, personality information related to the activity information and user personality information derived by the personality analysis means. Scoring is performed based on suitability, and both the activity information with a high score and the activity information with a low score are extracted and output.

  A program according to the present invention that achieves the above object includes a first step of acquiring information about a user, a second step of deriving the personality information of the user from the acquired information about the user, and at least the personality information and activity information of the user For each of the activity information stored in the activity information storage means stored in association with each other, scoring is performed based on the suitability with the personality information of the user derived in the second step, and the high score And causing the computer to execute a third step of extracting and outputting both the activity information and the activity information having a low score.

  The recommendation system configured as described above analyzes personality based on information about the user, and dares to extract activity information that does not match the personality, thus renewing the user's interest and bringing a new experience to the user be able to. In addition, since activity information matching the personality of the user is output, satisfactory results can be provided even for users who desire activity information within the range of their own interest.

  If the information acquisition means acquires information from the server of the social networking service registered by the user, the user only needs to input the minimum information to the recommendation system. The time and effort of operation can be minimized.

  The information acquisition unit acquires image data and text data relating to a user, and the personality analysis unit determines, based on an object extracted from the image data and the text data, for each element characterizing a preset personality. If the user personality information is derived by scoring, the user personality information is analyzed from both the image data and the text data, and thus the user personality information can be derived with higher accuracy.

  If the activity information extracting means extracts and outputs a predetermined number of the activity information in order from the highest score and a predetermined number of the activity information in order from the lowest score, the activity information extracting means outputs the most to the user. Matching information and information completely opposite to them are obtained, and a result that is surprising to the user can be output.

  If the activity information extraction means extracts and outputs four or less pieces of the activity information, it can prevent the user from being confused by the minimum necessary activity information.

  The activity information storage means stores location information in association with the activity information, the information acquisition means acquires location information transmitted by a user, and the activity information extraction means includes the activity information. If narrowing down is performed based on the position information, only information that matches the location where the user goes out can be obtained, and useless information can be prevented from being output.

The activity information storage means stores weather information in association with the activity information,
The information acquisition means acquires schedule information transmitted by the user and acquires weather information from the acquired schedule information. The activity information extraction means is based on the weather information for the activity information. If additional scoring is performed, activity information according to the weather can be provided.

  The program configured as described above analyzes personality based on information about the user and extracts activity information that does not match the personality, so that the user's new interest is aroused and a new experience for the user is brought about. Can do. In addition, since activity information matching the personality of the user is output, satisfactory results can be provided even for users who desire activity information within the range of their own interest.

It is a block diagram of the whole system containing the recommendation system in this embodiment. It is a flowchart of the recommendation information output by a recommendation system. It is explanatory drawing which shows the structure of the information memorize | stored in the activity information storage means. It is explanatory drawing which shows the process which sorts and extracts the scored activity information.

  Embodiments of the present invention will be described below with reference to the drawings. The system of the present embodiment is configured to provide information on activities at a travel destination in response to a request from a user.

  FIG. 1 shows a configuration diagram of the entire system including a recommendation system 10 in the present embodiment. As shown in this figure, the system of this embodiment includes a recommendation system 10, a processing system 11 that performs various processes in response to requests from the recommendation system 10, an SNS server 12 that stores information about the user, And a user terminal 13 to be used. The recommendation system 10 can communicate between the processing system 11, the SNS server 12, and the user terminal 13.

  The recommendation system 10 includes information acquisition means 20 for acquiring information such as a travel schedule and information about the user, a personality analysis means 21 for analyzing the personality of the user based on the information about the user, information such as a travel schedule, Activity information extracting means 22 for extracting and outputting recommended activity information from information about the user.

  The processing system 11 evaluates personality elements from information necessary for personality analysis, image analysis means 30 that extracts and outputs objects included in an image, weather information provision means 31 that outputs weather information from date information, and the like. The personality element evaluation means 32 for outputting and the activity information storage means 33 for storing the activity information in association with other information. Functions by these means are provided as an API (Application Programming Interface) of the processing system 11, and the recommendation system 10 can call and use these functions.

  Each means constituting the recommendation system 10 and the processing system 11 can be realized by a computer having a central processing unit and a memory. The computer may be a server device or a small computer such as a PC.

  The user terminal 13 can be configured by a portable information terminal such as a smartphone or a computer such as a PC. An application for using this system is installed in the user terminal 13 in advance. In this application, input of information to be transmitted to the recommendation system 10, display of activity information received from the recommendation system 10, reservation of activities, and the like can be performed. Note that the user terminal 13 may exchange information with the recommendation system 10 using a general-purpose application such as a browser instead of a dedicated application.

  The SNS server 12 stores text and image data posted by the user. The user terminal 13 is a portable information terminal such as a smartphone, and can perform wireless communication with the recommendation system 10.

  Next, the detailed operation of each unit will be described while explaining the flow of the recommendation system 10 of the present embodiment. FIG. 2 shows a flowchart of recommended information output by the recommendation system 10. First, the user inputs a schedule and place for performing an activity on the user terminal 13 (S1).

  The schedule and location information input at the user terminal 13 is transmitted to the recommendation system 10. The recommendation system 10 that has received the information first collects information about the user. For this purpose, the information acquisition means 20 of the recommendation system 10 acquires information about the user from the SNS server 12 (S2). Acquisition of information from the SNS server 12 is performed through user consent and authentication.

  From the SNS server 12, text data and image data posted by the user are acquired. Further, the image data is analyzed to extract an object (S3). In the analysis of image data, an object included in an image, a scene such as the sea or snow, a place of a landscape, a person, and a person's facial expression are extracted as objects. For the analysis of the image data, a known method such as a statistical method or a machine learning method such as deep learning can be used.

  The recommendation system 10 sends the image data acquired from the SNS server 12 to the processing system 11, and the processing system 11 extracts an object of the image data in the image analysis means 30. The processing system 11 transmits the object information of the extracted image data to the recommendation system 10.

  Next, the recommendation system 10 uses the image data object and text data acquired from the SNS server 12 to perform the personality analysis of the user by the personality analysis means 21 (S4). The personality analyzing means 21 shapes the above-mentioned image data object and text data into a predetermined format, and transmits them to the personality element evaluating means 32 of the processing system 11. The personality element evaluation means 32 analyzes the personality of the user based on the five characteristics and the subordinate elements of the characteristics using the big five theory. The five characteristics evaluated by the personality element evaluation means 32 are cooperation, integrity, extroversion, emotional instability, and openness.

  The personality element evaluation means 32 is preliminarily input with the relationship between the image data object and text data and each of the above-mentioned elements, and based on this, the image data object analyzed at S3 and the text data acquired at S2 are obtained. Evaluate and score for 5 properties. Note that the relationship between the object or text data of the image data and each of the above-described elements is not limited to those input in advance, and may be acquired by, for example, machine learning. Users whose score distributions calculated by the personality element evaluation means 32 are close to each other can be evaluated as having close personalities, and users who have completely different score distributions can be evaluated as having low personalities.

  The personality element evaluation unit 32 that has evaluated the personality characteristics transmits the personality information that is the result of the personality element evaluation unit 32 to the personality analysis unit 21 of the recommendation system 10.

  When the user's personality analysis is performed, the recommendation system 10 acquires weather information (S5). The information acquisition unit 20 transmits the schedule and location information input by the user in S1 to the processing system 11, and the processing system 11 outputs the weather information on the schedule and location in the weather information providing unit 31. Is transmitted to the recommendation system 10. The weather information includes weather information such as sunny, rainy, and snow, and weather observation data such as temperature, wind direction, wind speed, humidity and pressure, dew point, visibility, UV index, and precipitation probability.

  Next, the recommendation system 10 scores the activity information based on the personality information of the user analyzed in S4 (S6). The activity information extraction unit 22 performs scoring on the activity information stored in the activity information storage unit 33 of the processing system 11.

  FIG. 3 shows the structure of information stored in the activity information storage means 33. The activity information storage means 33 stores a plurality of activity information such as activity information 1, 2,. Each activity information is contents of activities such as cycling, diving, skiing, and the like. In each activity information, personality information of a person related to the activity is stored in association with each other. A plurality of user personality information is stored for each activity information. The personality information of the person related to the activity is collected in advance and stored in the activity information storage means 33. For example, information on a bulletin board on the Internet is accumulated, and writing related to an activity is extracted from a large number of writings. Then, personality information of the person who performed the writing is acquired from the extracted writing text data. The personality information can be acquired by inputting the extracted written text data into the personality analysis means 21 used in S4. The personality information may be collected by other methods. The acquired personality information is associated with the activity information corresponding to the activity related to the original writing, and is stored in the activity information storage means 33.

  Further, each activity information stores information on a business provider that provides the activity in association with each other. The information on the business includes information on the business name and the location where the activity is provided. A plurality of information on the business operators is stored for each activity information.

  Furthermore, weather information suitable for the activity is stored in association with each activity information. The weather information includes information such as the weather suitable for the corresponding activity, temperature, and wind speed.

  In S6, the activity information extraction unit 22 compares the personality information of the user analyzed in S4 with the personality information stored in the activity information storage unit 33, and performs scoring for each activity information. In scoring, a high score is given to activity information associated with personality information close to the personality information of the user analyzed in S4, and a low score is assigned to activity information associated with personality information far from the personality information of the user analyzed in S4. wear. Various methods are conceivable. For example, for all personality information associated with certain activity information, the distribution difference from the personality information analyzed in S4 is evaluated and scored, and the total score is given. Is calculated. The same total score is calculated for each activity information and used as a score for each activity information.

  Next, the recommendation system 10 narrows down the activity information based on the location information input by the user (S7). The activity information extraction unit 22 collates the information on the place input by the user with the activity providing location included in the business operator information stored in the activity information storage unit 33, and determines the activity that can be provided at the location input by the user. Extract the corresponding activity information. Here, activities that are not provided at the location entered by the user are eliminated.

  Next, the recommendation system 10 performs additional scoring based on the weather information acquired in S5 (S8). The activity information extraction means 22 corrects the score attached at S6 based on the weather information for the activity information narrowed down at S7. The weather information is stored in association with each activity information in the activity information storage means 33 as described above. The activity information extracting means 22 compares the weather information acquired in S5 with the weather information associated with the activity information, and increases the score of the activity information when both are close, and when both are far, Lower the activity information score.

  Through the step S8, a score for each piece of activity information is calculated. As shown in FIG. 4, the activity information extraction unit 22 arranges each activity information in the activity information storage unit 33 in the order of the calculated score, and the activity score is extracted from the highest score and the lowest score. Are extracted respectively (S9). In the example of FIG. 4, a total of four pieces of activity information are extracted: activity information 12 and activity information 3 with a high score, activity information 9 with a low score, and activity information 4. Then, the activity information extraction means 22 outputs the extracted activity information together with corresponding business operator information (S10). This result is transmitted to the user terminal 13.

  The recommendation system 10 stores a program for causing a computer to execute these steps.

  Thus, the recommendation system 10 of this embodiment scores activity information with a user's personality information, and outputs both activity information with a high score and activity information with a low score as a result. Activity information with a high score matches the personality of the user and is displayed as a recommended activity. On the other hand, activity information with a low score is something that the user does not normally come up with, and is displayed as a surprise activity. The user obtains both a result matching his personality and a result matching his personality.

  Since there are a wide variety of activities, from indoors to outdoors, the activities that do not match the user's personality are intentionally output as a result, so that new interest can be generated and a new experience for the user can be brought about. In addition, since an activity that matches the user's personality is output, a satisfactory result can be provided even for a user who desires an activity within the range of his / her interest.

  In the present embodiment, the activity information output by the recommendation system 10 includes a total of four activity information, two with a high score and two with a low score. The number of pieces of activity information to be output can be arbitrarily set. However, if the number of pieces of activity information to be output increases, it becomes difficult for the user to select, so the number of results to be output is preferably four or less.

  The activity information provided to the user from the recommendation system 10 includes information on a business operator that provides the activity, and reservations can be made from the user terminal 13. The service provider of the recommendation system 10 can obtain a certain amount of revenue in accordance with the reservation made by the user to the business operator.

  Since the recommendation system 10 obtains information about the user from the SNS server 12 of the social networking service registered by the user, the user only needs to input the minimum information to the recommendation system 10. Can be minimized.

  Note that the present invention is not limited to the above-described embodiments, and various modifications can be made by those skilled in the art within the technical idea of the present invention. Hereinafter, modifications of the present invention will be described.

  In the above-described embodiment, the user inputs the place where the activity is performed. However, the place may not be specified. In this case, since the user cannot predict in advance where the activity will be proposed, the output result can be surprising.

  Moreover, in the above-mentioned embodiment, although the information regarding a user is acquired from the SNS server 12, you may acquire the information regarding a user by other than that. For example, the user may be allowed to input information regarding personality, hobbies, and the like.

  In the above-described embodiment, the recommendation system 10 and the processing system 11 are separated, and the recommendation system 10 uses the API provided by the processing system 11, but the recommendation system 10 uses each means of the processing system 11. And the system may be an integrated system.

  In the above-described embodiment, as a result of scoring, a predetermined number of information is extracted in order from the highest score, and a predetermined number of information is extracted in order from the lowest score. The information extraction method is not limited to this. For example, in the extraction of information with a low score, the information with the highest or lowest score is included, such as extracting the information with the second lowest score and the information with the third lowest score. It may not be possible.

  In the above-described embodiment, the activity information extraction unit 22 performs additional scoring on the activity information based on the weather information, but may output the result without using the weather information. .

  The recommendation system 10 can also be applied to activity information other than activities at travel destinations. For example, the recommendation system 10 of the present invention can be configured using activity information such as restaurants, apparel shops, temple visits, cooking recipes, and the like.

DESCRIPTION OF SYMBOLS 10 Recommendation system 11 Processing system 12 SNS server 13 User terminal 20 Information acquisition means 21 Personality analysis means 22 Activity information extraction means 30 Image analysis means 31 Weather information provision means 32 Personality element evaluation means 33 Activity information storage means

Claims (8)

Information acquisition means for acquiring information about the user;
Personality analysis means for deriving personality information of the user from information acquired by the information acquisition means;
An activity for extracting and outputting the activity information from the activity information storage means in which at least the activity information and the personality information related to the activity information are stored in association with each other based on the user personality information derived by the personality analysis means An information extraction means,
The activity information extraction means is based on the suitability between the personality information related to the activity information and the user personality information derived by the personality analysis means for each of the activity information stored in the activity information storage means. A recommendation system characterized in that scoring is performed, and both the activity information with a high score and the activity information with a low score are extracted and output.   The recommendation system according to claim 1, wherein the information acquisition unit acquires information from a server of a social networking service registered by the user. The information acquisition means acquires image data and text data related to a user,
The personality analysis means derives the personality information of the user by performing scoring for each element characterizing a predetermined personality based on the object extracted from the image data and the text data. The recommendation system according to claim 1 or 2.   2. The activity information extracting means extracts and outputs a predetermined number of the activity information in order from the highest score and a predetermined number of the activity information in order from the lowest score. The recommendation system of any one of -3.   The recommendation system according to any one of claims 1 to 4, wherein the activity information extraction unit extracts and outputs four or less pieces of the activity information. The activity information storage means stores location information in association with the activity information,
The information acquisition means acquires information on a position transmitted by the user,
The recommendation system according to claim 1, wherein the activity information extraction unit narrows down the activity information based on the position information. The activity information storage means stores weather information in association with the activity information,
The information acquisition means acquires information on a schedule transmitted by the user, acquires weather information from the acquired schedule information,
The recommendation system according to claim 1, wherein the activity information extraction unit performs additional scoring on the activity information based on the weather information. A first step of obtaining information about the user;
A second step of deriving personality information of the user from the acquired information about the user;
For each of the activity information stored in the activity information storage means stored in association with at least the user personality information and the activity information, the score is based on the compatibility with the user personality information derived in the second step. Performing a ring to extract and output both the activity information with a high score and the activity information with a low score; and
A program that causes a computer to execute.

Images