KR102096419B1 - Travel companion and product providing system based on location information and big data analysis and the method using it - Google Patents

Abstract

The present invention relates to a system and method for providing travel companions and products based on location information and big data analysis. More particularly, the present invention relates to a system and method for providing travel companions and products based on location information and big data analysis which provide various services such as ″my own travel mate″, ″finding a companion″, ″applying for companion″, ″making local friends″, and ″the more you gather, the more enjoyable travel service″, so that it is easy and safe to find a companion to travel with.

Description

Travel companion and product providing system based on location information and big data analysis and the method using it}

The present invention relates to a system and method for providing travel companions and products based on location information and big data analysis, and more specifically, providing a user's preferred travel and travel product recommendations, finding companions by region around the world, and finding companions adjacent to me Therefore, it is related to a system and a method for helping a user to conveniently and safely obtain a companion to travel together.

With the recent development of transportation technology and increasing interest in leisure life, travel has become one of the major leisure activities.

In order to enjoy a highly satisfactory trip, people collect information related to the destination they visit, consult with local residents or travel experts, or even recommend a travel agency package schedule.

However, these tasks are cumbersome and often fail to reflect individual characteristics, resulting in unsatisfactory results.

In order to solve these problems, a lot of research has been conducted related to an automated recommendation service that identifies a user's tendency and recommends a travel itinerary.

However, research on personalization services is insufficient because most of the studies apply common criteria for all users.

Commonly used techniques for collecting personal information and analyzing collected information to recommend products are collaborative filtering and content-based filtering. As a technique, a method of recommending using user's product purchase history and user's similarity is used. However, this method cannot make a recommendation preferred by the user when the data is not sufficiently secured.

In addition, due to the nature of the overseas travel market, it is a pattern that consumes travel about once a year, and because there are almost no travels to the same area, it is difficult to apply to the travel market with general product recommendation systems such as collaborative filtering and content-based filtering. There are limits.

In addition, the conventional clustering technique does not reflect the preference of the recommended user, and has the disadvantage of recommending according to the preferences of other users.

In addition, since the pattern matching technique performs recommendation by comparing the type and distance of the point of interest (POI) with other users, there is a disadvantage that the detailed preferences of the user cannot be contradicted.

Accordingly, the present invention proposes a preferred personalized recommendation system (algorithm) for each user by applying a combination of a cooperative filtering, a content-based filtering technique, and a travel recommendation agent of the present invention to solve this problem.

In addition, it is becoming increasingly difficult to meet the itinerary with family and friends as well as single-person households that grow rapidly every year. There is a need.

In addition, there is a need for a service that can solve the problems of single travelers who are hesitant when traveling alone, and can provide travelers with more valuable experiences as well as exchange with various people.

Korean Patent Publication No. 2014-0087175 Korean Registered Patent No. 0907744 Korean Patent Publication No. 2002-0033202 Korean Patent Publication No. 2003-0003396

The present invention has been made to solve the above problems, and the travel propensity and travel purpose preferred by the solo traveler, such as the gender and age of the solo traveler, the preferred region or country, the preferred travel date, the preferred product, and the purchase price. The purpose is to provide a travel companion and a product providing system based on location data and big data analysis that can build related big data.

In addition, the present invention is a travel based on location information and big data analysis that can recommend travel products that can be enjoyed together or accompanied by a traveler in a suitable location according to the traveler's real-time location, travel propensity, travel purpose, and travel theme. The purpose is to provide a companion and product delivery system.

In order to solve the above problems, the present invention is a processing server for recommending a companion to a user and a travel product that can be enjoyed with the companion; A database server that stores the user information, companion information, and travel product information; It includes; a mobile terminal for receiving and displaying a travel product that can be enjoyed with the companion from the processing server, and receiving additional companion or travel-related input information and transmitting it to the processing server.

The processing server is composed of a travel recommendation agent module including a travel recommendation algorithm and a travel product recommendation algorithm that recommend the travel companion and the travel companion, and to reflect the preferences of the user and the individualized travel preferences, A user clustering module for assuming that there is a correlation between similar groups through user's interest destination search information classified by gender and age, and recommending a tour in a region having a high number of views by similar groups; The travel recommendation data generated and recommended by the travel recommendation agent and the user clustering module are stored in a database, the individual travel preferences are modeled through the data mining technique, and satisfaction survey information and travel destination search after travel by each classified user group Includes a travel analysis and clustering module that analyzes a user's travel propensity and companion preferences by providing weighted information to provide customized recommendations for the next travel destination.

The travel analysis and clustering module generates a conversion ratio of travel products repurchased by providing customized recommendations for the next travel destination by assigning weights to satisfaction survey information and travel destination search information for each user group after travel.

The processing server further includes a collaborative filtering algorithm for recommending to other users, but the collaborative filtering algorithm filters and extracts the top recommended content candidates having the highest user preference scores and classifies them into a plurality of categories to distribute travel propensity. Is displayed.

The present invention provides a method for providing travel companions and product services through a processing server of a travel companion and product providing system based on location information and big data analysis, collecting user demographic information and storing it in a database server (S101); Filtering from the user demographic information to valid information calculated by a collaborative filtering algorithm for recommending to other users (S102); Collecting users' real-time location data and storing them in a database server (S103); Recommending a travel companion and a travel product through the valid information (S104); Providing a customized recommendation for the destination to generate a conversion rate for repurchase of the travel product (S105); Calculating a weight by referring to the conversion ratio (S106); Analyzing a user's travel propensity and companion preferences (S107); Analyzing a user's travel propensity and companion preferences and filtering to recommend companion and travel information (S108); And recommending new companion and travel information based on the information filtered in step S108 (S109).

The method further includes filtering and extracting the candidates with the highest recommendation content having the highest user preference score through the collaborative filtering algorithm and classifying them into a plurality of categories to display a travel propensity distribution.

The method further includes constructing travel preference data generated through the national statistics portal server and the travel statistics portal server and modeling individual travel preferences through a data mining technique.

Through the travel analysis and clustering module of the travel companion and product providing system based on the location information and big data analysis, weighting is applied to satisfaction survey information and travel destination search information for each user group to provide customized recommendations for the next travel destination. And generating and storing the travel product repurchased conversion ratio in a database server.

The present invention made as described above does not fit the schedule with a friend or family, and when you need to travel alone, when you want to find someone who can do what you are interested in when traveling, travel to the city where you live or the city where you are traveling It is easy to use when people are curious.

In addition, according to the present invention, a statistically predictable recommendation is provided through demographic consumption analysis by analyzing gender, age, occupation, etc. of other users and conventional clustering techniques and pattern matching techniques to easily and safely find a companion to travel with. This is possible.

In addition, according to the present invention, as user-based travel data accumulates according to gender, age, region or country, and interest, it is possible to accurately reflect tastes by filtering user-based data.

In addition, the present invention can reflect the exact taste by a filtering system that recommends appropriate information about the next preferred area after checking the travel history through a companion recommendation algorithm and the like in order to analyze a user's travel preference and companion preference.

In addition, according to the present invention, a more precise recommendation is made than a method of recommending simply using local information by analyzing a region viewed by user in correspondence with a date.

In addition, according to the present invention, after collecting and assigning weights to satisfaction data, when recommendation to a user decreases the purchase conversion rate, the weight is decreased, and the high-weighted data is re-recommended to increase satisfaction to increase the purchase conversion rate.

In addition, the present invention allows more precise recommendations to be made by grouping people with similar preferences through calculating the similarity, or using a method of recommending to a person who has not been to a specific region.

1 is a view showing in detail the configuration of a travel companion and a product providing system based on location information and big data analysis according to the prior invention.
2 is a flowchart showing a travel companion and a product providing method based on location information and big data analysis according to an embodiment of the present invention.
FIG. 3 is a view showing a menu for “registering companion”, “viewing companion profile”, and “chat” service according to another embodiment of the present invention.
4 is a view showing a menu for the "My Own Travel Mate", "Browse My Accompaniment", and "Apply for Travel" services according to an embodiment of the present invention.
FIG. 5 is a view showing a menu for providing a “making friends in the neighborhood” and “a more pleasant trip together” service according to another embodiment of the present invention.
6 is a flowchart showing a method of providing travel companions and product services through a processing server of a travel companion and product provider system based on location information and big data analysis according to another embodiment of the present invention.

In order to fully understand the present invention, preferred embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments of the present invention may be modified in various forms, and the scope of the present invention should not be interpreted as being limited to the embodiments described in detail below. This embodiment is provided to more fully describe the present invention to those skilled in the art. Accordingly, the shape of the elements in the drawings may be exaggerated to emphasize a clearer description. It should be noted that in each drawing, the same members may be indicated by the same reference numerals. In addition, detailed descriptions of well-known functions and configurations that are judged to unnecessarily obscure the subject matter of the present invention are omitted.

As shown in FIG. 1, the present invention comprises a mobile terminal 10, a database server 20, a processing server 30, and the like.

The mobile terminal 10 receives and displays travel products that can be enjoyed with the companion from the processing server 30, receives the added companion or travel-related input information, and transmits it to the processing server 30 to be.

The database server 20 stores the user information, companion information and travel product information

The processing server 30 is a server that recommends a companion to the user and a travel product that can be enjoyed with the companion to the mobile terminal 10.

Specifically, the processing server 30 is a travel recommendation agent 31, a user clustering module 32, and a travel recommendation agent 31 is accompanied by an algorithm that recommends a companion recommendation product recommending a companion and a companion. It includes a module and a travel recommendation module including an algorithm for recommending travel products that can be enjoyed together with the companion recommended in the companion recommendation module.

The user clustering module 32 assumes that there is a relationship between similar groups and searches for similar groups through search information of interest destinations of users classified by gender and age, in order to reflect the user preferences and the individualized travel preferences It is a module that recommends travel in high-density areas.

The similarity calculation process of the similar group analyzes the companion and travel usage information, and selects two companions and travel usage selectable among two or more companions and travel usage information continuously used by the same user within a preset time interval. The information combination may be determined as a similar pair, and a time weight inversely proportional to an interval between use points used by the same user may be applied to the similar pair.

The travel analysis and clustering module 33 stores the travel preference data recommended and generated through the travel recommendation agent 31 and the user clustering module 32 in a database, and models individual travel preferences through data mining techniques. , By assigning weights to the satisfaction survey information and travel destination search information for each classified user group, it provides customized recommendations for the next travel destination to analyze the user's travel propensity and companion preferences.

The travel analysis and clustering module 33 generates a conversion rate of repurchase of travel products to provide customized recommendations for the next travel destination by assigning weights to satisfaction survey information and travel search information for each user group after travel. do.

The conversion rate of the repurchase of the travel product is generated by processing the travel product based filtering method based on the demographic information of the user's gender and age information.

The travel analysis and clustering module 33 filters and extracts top recommendation content candidates having the highest user preference scores through a collaborative filtering algorithm for recommending to other users, classifies them into a plurality of categories, and classifies travel propensity distributions into mobile terminals ( 10).

Specifically, as shown in FIG. 3, when the companion registration service is started, in the present invention, the user receives an application user's membership registration in order to subdivide the user preferences using a user-based collaborative filtering method. It recommends destinations to other users based on the user's gender and age information, which is statistical information, and based on the detailed user's travel history information through content-based filtering.

However, there is a limit to consistently classifying the user's preferences only because the gender and age of the user and companion are the same, and there is a limitation in recommendation even when sufficient travel history data is not accumulated (cold-start problem).

 In order to solve these problems, the present invention obtains and classifies individual travel propensity information, and sufficiently reflects the user's travel preference based on the preferred travel theme and preferred companion data for each travel destination, and solves the problem caused by insufficient data. Can be solved.

Therefore, the present invention can solve the problem by reflecting the personalized travel propensity of each individual in that the preference of the recommended user, which is a disadvantage of the conventional clustering technique, is not reflected and the companion and the travel product are provided.

The cold-start problem is solved by assuming that there is a correlation between similar groups through user's interest destination search information classified by gender and age of the companion information, and recommending a travel in a region with a high number of similar groups.

For example, travel preference data generated through a national statistics portal server, a travel statistics portal server, and the like are constructed, and individual travel preferences are modeled through data mining techniques.

As a technique for analyzing the association between the travel preference data downloaded through a national statistics portal server, there are techniques for analyzing support, confidence, and lift.

That is, n is the total number of user sessions, n (A) is the number of sessions that have visited the travel product page A, n (A, B) is the number of sessions that have visited both A and B, n (A-> B) Is defined as the number of sessions that visited A first and then B.

Support (A-> B) = n (A, B) / n

Confidence (A-> B) = n (A, B) / n (A)

Sequence (A-> B) = n (A-> B) / n (A)

The content-based filtering method is a method of recommending travel products having similar characteristics through profiling of travel products purchased in the past, so profiling of travel products becomes relatively important and travel products In this case, the disadvantages of the collaborative filtering recommendation method due to lack of data may be offset.

The rule-based filtering recommendation method approaches based on the existence of a specific rule in the user's purchase pattern.

This rule-based short-term recommendation time can solve Cold-Start problems, but is not suitable for reflecting personal profiles.

The user profiling method through web log analysis (Web Usage Mining) is a common method (login, member purchase), and is often used in cases where user information cannot be specified.

A Priori Algorithm (A Priori Algorithm) is an algorithm that generates a series of rules that tell which two sets of items occur frequently.

The location data of users is collected to form a semantic category group for each POI, and a detailed group is formed according to the distribution of residence time.

라고 할 때, 각 POI에 대한 사용자의 선호도를 계산하기 위해 POI들의 시맨틱 정보를 추출한다. It also creates user profiles for each user. First, the entire set of POIs

In order to calculate the user preference for each POI, semantic information of the POIs is extracted.

Based on the extracted information, a semantic category vector for each POI is generated, and groups of POIs having the same meaning are formed.

A semantic category group is a collection of places with the same semantic category.

인 POI 집합(C 는 전체 시맨틱 카테고리 집합)이다.In other words, the semantic category

Is the POI set (C is the entire semantic category set).

은 시맨틱 카테고리 그룹

의 세부 그룹이며, 체류시간 분포가 같은 장소들의 집합으로 다음과 같은 성질을 지닌다.Residence time distribution group

Semantic Category Group

It is a subgroup of and is a collection of places with the same residence time distribution and has the following characteristics.

라고 할 때, 사용자 u가 방문한 적 없는 장소 p'의 예상 체류시간

는 p'와 같은 체류시간 분포 그룹에 속하는 장소(

)들에서의 평균 체류시간이다.All the places you visit

Is the estimated residence time of the place p 'where user u has never visited

Is a place belonging to the residence time distribution group such as p '(

).

The objective function Γ for the user's satisfaction with the travel itinerary H is the sum of the user's semantic category preferences taking into account the difference between the expected residence time and the actual residence time for the POI belonging to H.

When returning the Top-K schedule from the online module, the user's preference vector is the basis, so the preference function must be calculated in real time.

If the preference function is calculated for the entire travel itinerary related to the destination, there are too many candidate schedules. Therefore, a technique for calculating the preference function for candidate itineraries in a short time and efficiently constructing a personalized Top-K schedule is needed. Do.

Threshold Algorithm (TA) is an algorithm developed to efficiently return the Top-K schedule in applications that use multidimensional spatial vectors.

However, due to the nature of the Threshold Algorithm (TA), it can be applied only when the objective function has a property of continuously increasing. Therefore, in the proposed system, the minimum range function with the monotonic increase function property is set and used as follows.

"는 일정이 포함하고 있는 모든 시맨틱 카테고리를 표현하는 벡터로 추천 일정에 포함된 모든 장소의 시맨틱 카테고리 벡터의 합으로 표현된다.Recommended Schedule Limited Vector "

"Is a vector representing all semantic categories included in the schedule, and is expressed as the sum of the semantic category vectors of all places included in the recommended schedule.

"는 여행체류시간에 대응하는 선호도를 제외하고 추천 일정의 시맨틱 카테고리에 대해서만 고려한 사용자의 선호도이다.Limited functions for recommended itineraries "

"Is the preference of the user considering only the semantic category of the recommended schedule except the preference corresponding to the travel time.

Create a table composed of <iti_id, tag_id, weight> using the minimum range function.

Here, iti_id and tag_id represent unique numbers for the schedule and semantic categories, respectively, and weight is a vector component value of the minimum range function for the semantic category for each schedule.

을 생성한다. Sort list independently constructed for each vector dimension by partitioning the least-range function table

Produces

When a unit vector for n semantic categories is given as follows, the query processing algorithm is as follows.

)의 경쟁력(CV) 측정을 위하여, 월간 검색 량(AS)과 블로그 개수(NB), 상위 노출 블로그 10건의 평균 작성 일과 현재 날짜와의 차이(ADB), 연관성(TF-IDF)을 고려하여 아래 수학식 7과 같이 계산하였다. 연관 키워드들은 계산된 경쟁력의 값 역순으로 추천된다.The present invention extracts and filters each keyword in the associated keyword group (

To measure the competitiveness (CV) of the), considering the monthly search volume (AS) and the number of blogs (NB), the average creation date of the top 10 blogs and the difference between the current date (ADB) and relevance (TF-IDF), It was calculated as in Equation 7. Related keywords are recommended in the reverse order of the calculated competitive value.

)가 포털 시스템에서 최근 한 달 이내에 검색된 횟수를 의미한다.Monthly search volume (AS) is the relevant keyword (

) Means the number of times the portal system was searched within the last month.

)의 포털 시스템 검색 결과 포함된 블로그의 전체 개수를 나타내며, ADB는 해당 연관 키워드(

)의 포털 시스템 검색 결과 The number of blogs (NB) is the relevant keyword (

) Indicates the total number of blogs included in the portal system search results, and ADB indicates the relevant keyword (

) Portal system search results

This is the average value calculated by calculating the difference (unit: day) between the top 10 exposed blogs and the current date. The association (TF-IDF) value uses the calculated value when extracting the relevant keyword.

Even if a keyword with low relevance is competitive, it is not useful for the intention of posting. Therefore, keywords with high relevance can be included in the recommended keyword group.

The processing server 30 calculates interest in the category of the behavior attribute information extracted using time weights and user interest weights, and sums interests on the same travel product.

At this time, the user himself / herself may select a desired travel category in advance and input weights only to the selected travel category, or if he / she does not have the desired travel category, create it by directly inputting it, or may be provided in a preference index based user matching system Use the category classification system.

When the appointment time interval is within a predetermined minimum appointment time interval in a specific travel area and the maximum number of appointments is 3 times, the processing server 30 transmits a specific travel product more than the maximum number of appointments within the minimum appointment time interval for the same travel product. It is recommended as a group of companions by grouping companions showing the form of consumption.

By applying weights according to the user's behavior and / or usage time information, the travel product preference for each user is calculated for each travel product.

For example, if 10 users have travel product A and travel product B used together on the same day 30 days ago, and 10 users have travel product C 10 days ago and travel product D 1 day ago. Although it happened more recently, since the travel product A and the travel product B occurred on the same day, the weight of the travel product A and the travel product B is calculated to be more similar to the travel product C and the travel product D.

Therefore, in the present invention, as the data is accumulated as the individual preferences are subdivided, the exact taste can be reflected, and after the trip is categorized, the satisfaction survey information and the travel destination search information are weighted to provide customized recommendations for the next travel destination. Increase the product conversion rate.

For example, from the city gourmet tours of Figures 4 and 5 to activities, the "My Own Travel Mate", "Search for Companion" by continent / city, and "Apply for Companion" by hobby / theme service By subdividing propensity and assigning weights, it is possible to increase the conversion rate of travel products by providing customized recommendations for the next destination.

In other words, for each individual who has applied for “My Own Travel Mate”, “Search for Companion”, and “Apply for Companion” service products, the weighted product is recommended first for each city, by continent, and for each hobby theme. do.

As another embodiment of the present invention, the processing server 30 is based on a semantic web service to provide a travel plan recommendation system (provided by a friend of the neighborhood, a gathering of friends) in order to provide a service of “making a friend in the neighborhood” and “making a travel more enjoyable” in FIG. Recommendations, etc.).

If you request a companion ticket, enter your vacation schedule and personal preferences to a travel agent who can propose a travel plan recommendation system based on the semantic web service, so you can easily create and participate in all the local friends and meetings required for the trip. I can apply.

For example, when requesting a companion application, the similarity according to the inversion, omission, insertion, or modification of words used in the user profile can be measured using a POI search technique.

This is because the words used when traveling in different neighborhoods or regions or countries do not match exactly.

In addition, as a result of a search for a query to search for a companion, if there are multiple search words, the greater the degree of matching between the order between syllables in the first syllable pair constituting the search word and the second syllable pair in the second syllable pair constituting the query word, the higher the priority. By outputting it, it is possible to more easily recognize the search word in which the syllable is arranged closest to the syllable arrangement of the query.

In addition, the present invention analyzes a region that is frequently viewed by each user according to a date, so that a more precise recommendation is made than a method simply recommending using region information.

In addition, the present invention focuses on safety of companions by frequently monitoring whether a weight or an output score (or purchase conversion rate, preference score) causes a rapid change (eg, EMA change rate) at a certain time or by subdividing individual propensity. I can put it.
For example, the processing server may continuously monitor the user's preference, travel preference, and EMA fluctuation rate, which calculates the rate of change over a certain period of time for travel preferences and companion preferences, to detect when the rate of change is greater than or equal to a certain value.

In addition, according to the present invention, if there is no desired travel category, a part that can be incorrectly classified in a predetermined category classification system that can be generated by directly inputting it or generated by a preference index based user matching system, such as Naive Bayes and Support Vector Machine, etc. It can also be corrected through.

In addition, Naive Bayes and Support Vector Machine can analyze and block words used in companion application to identify and block attempts to hide male / female / age / region in advance.

Hereinafter, a method of providing travel companion and travel product service through a processing server of a travel companion and product providing system based on location information and big data analysis for implementation of the present invention will be described in detail with reference to FIG. 7.

First, user demographic information is collected and stored in a database server (S101).

Then, the user demographic information is filtered with valid information calculated by a collaborative filtering algorithm for recommending to other users (S102).

Continuing to collect real-time location data of users and store it in a database server (S103), recommend travel companions and travel products through the valid information (S104), and provide custom recommendations for the destinations to repurchase travel products By generating a ratio, the weight is calculated with reference to the conversion ratio (S106), and a user's travel propensity and companion preference are analyzed (S107).

The user's travel propensity and companion preference are analyzed and filtered to recommend companion and travel information (S108).

The new companion and travel information are recommended based on the information filtered in step S108 (S109).

Through the collaborative filtering algorithm, the top recommendation content candidates having the highest user preference scores are filtered and extracted and classified into a plurality of categories to display a travel propensity distribution.

In addition, the travel preference data generated through the national statistics portal server and the travel statistics portal server are constructed, and individual travel preferences are modeled through data mining techniques.

Through the travel analysis and clustering module of the travel companion and product provision system based on the location information and big data analysis, weighting is applied to the satisfaction survey information and travel search information for each user group to provide customized recommendations for the next travel destination. Generate conversion rate of repurchase of travel products; and store them in a database server.

10: mobile terminal
20: database server
30: processing server

Claims (8)

A processing server for recommending companions and companions through a mobile terminal to a user, and recommending travel products that can be enjoyed with the companions;
A database server that receives and stores user information, companion information, and travel product information from a processing server;
It includes; a mobile terminal receiving and displaying a travel product that can be enjoyed with the companion from the processing server, and receiving additional companion or travel-related input information and transmitting it to the processing server.
The processing server,
A travel recommendation agent module including a companion recommendation algorithm for recommending companions and an algorithm for recommending travel products;
In order to reflect the preference of the user and the individualized travel propensity, it is assumed that there is a correlation between similar groups through user's interest destination search information classified by gender and age, and travels in an area where the number of similar group views is more than a certain value User clustering module to recommend;
The travel recommendation agent and the travel preferences generated and generated through the user clustering module are stored in a database, the individual travel preferences are modeled through the data mining technique, and the satisfaction survey information and travel destination search information after travel by classified user groups Includes a travel analysis and clustering module that analyzes a user's travel propensity and companion preferences by providing customized recommendations for next destinations by assigning weights.
The processing server continuously monitors the user's preferences, travel preferences, and travel preferences and companion preferences continuously to detect when the rate of change of the EMA calculated for a certain period of time is equal to or greater than a certain value. Travel companion and product delivery system based on and big data analysis. delete According to claim 1,
The travel analysis and clustering module,
The travel group based on location information and big data analysis is characterized by generating a conversion rate of repurchase of travel products by providing weighted weights to the satisfaction survey information and travel destination search information for each user group, and then providing customized recommendations for the next travel destination. And product delivery system. According to claim 1,
The processing server further includes a collaborative filtering algorithm for recommending to other users,
Through the collaborative filtering algorithm, the top recommendation content candidate with the highest user preference score is filtered and extracted and classified into a plurality of categories to display the travel propensity distribution, and travel companions and products based on location information and big data analysis. Delivery system. In the method of providing travel companions and products based on location information and big data analysis through the processing server of claim 1,
Collecting user demographic information and storing it in a database server (S101);
Filtering from the user demographic information to valid information calculated by a collaborative filtering algorithm for recommending to other users (S102);
Collecting users' real-time location data and storing them in a database server (S103);
Recommending a travel companion and a travel product through the valid information (S104);
Generating a conversion rate of repurchase of travel products by providing a customized recommendation for the destination (S105);
Calculating a weight by referring to the conversion ratio (S106);
Analyzing a user's travel propensity and companion preferences (S107);
Analyzing a user's travel propensity and companion preferences and filtering to recommend companion and travel information (S108);
Recommending new companion and travel information based on the information filtered in step S108 (S109);
It includes the step of continuously monitoring to detect a case where the rate of change of the EMA, which calculates the rate of change for a certain period of time for the user's travel propensity and companion preference, is greater than or equal to a certain value.
Location information, characterized in that the user's preferences, travel preferences, and travel preferences and companion preferences are continuously monitored to detect when the rate of change of the EMA calculated for a certain period of time is greater than or equal to a certain value through a processing server And travel data and product delivery method based on big data analysis. The method of claim 5,
Filtering and extracting the top recommended content candidates having the highest user preference scores through the collaborative filtering algorithm and classifying them into a plurality of categories to display a travel propensity distribution; location information and big data analysis further comprising Based on how to accompany and offer products. The method of claim 5,
Based on the location information and big data analysis, characterized in that it further comprises the step of building travel preference data generated through the national statistics portal server and the travel statistics portal server and modeling the individual's travel preference through a data mining technique. How to accompany your trip and provide products. The method of claim 5,
Through the travel analysis and clustering module of the travel companion and product providing system based on the location information and big data analysis, after weighting the satisfaction survey information and search destination search information for each user group, the travel is provided by providing customized recommendations for the next destination The conversion rate of repurchasing the product; generating and storing in a database server; a travel companion and a product providing method based on location information and big data analysis.

Images