KR102313285B1 - Travel information providing system that provides contents selected by credibility estimation - Google Patents

Abstract

The present invention relates to a travel information providing system for providing content selected by reliability estimation, and more particularly, to a travel information providing system, comprising: an information collection module for collecting travel-related content including travel course and accommodation reviews; a model configuration module for constructing a reliability estimation model for estimating the reliability of the content based on the extracted keyword's importance, the author's evaluation level, and the recommendation level by applying machine learning to the collected travel-related contents; a content selection module for selecting travel-related content using the constructed reliability estimation model; and a content providing module that provides the selected travel-related content to travelers.
According to the travel information providing system that provides content selected by reliability estimation proposed in the present invention, travel selected using a reliability estimation model constructed by applying machine learning to travel-related content including travel course and accommodation reviews By providing relevant content, you can easily provide high-quality travel information to potential travelers to increase their satisfaction.
In addition, according to the travel information providing system that provides content selected by reliability estimation proposed in the present invention, the reliability of content is estimated based on the importance of keywords extracted from travel-related content, and the rating and recommendation of the author. By constructing and using a reliability estimation model for

Description

TRAVEL INFORMATION PROVIDING SYSTEM THAT PROVIDES CONTENTS SELECTED BY CREDIBILITY ESTIMATION

The present invention relates to a system for providing travel information, and more particularly, to a system for providing content selected by reliability estimation.

As leisure time increases and a lot of information about destinations is shared through active use of social media, travel is gaining popularity as one of leisure activities.

Accordingly, in order to collect information about the area to be traveled or the surrounding area, a travel course is planned by searching a website, and along the planned travel course, various experiences such as sightseeing, food, and shopping can be experienced by walking or using a vehicle. . However, the information on the Internet is so extensive that it takes a lot of effort for users to find the travel information they want.

Recently, interest in long-term travel, such as living for a month rather than short-term travel, is increasing, and in the case of long-term travel, there are many considerations compared to short-term travel, so it is more important to collect reliable information.

However, unverified and indiscriminate contents such as portal sites and social media provide information with low accuracy or no practical value to potential travelers who are planning a trip, and sometimes provide advertising information that is not directly experienced. As such, many travelers have difficulties in collecting information, so it is necessary to develop a technology to provide high-quality travel information by selecting the content itself from among a large amount of content.

On the other hand, as prior art related to the present invention, Patent Publication No. 10-2019-0093774 (title of invention: travel information providing system through Internet information analysis, publication date: August 12, 2019) has been disclosed.

The present invention has been proposed to solve the above problems of the previously proposed methods, and travel-related content selected using a reliability estimation model constructed by applying machine learning to travel-related content including travel course and accommodation reviews An object of the present invention is to provide a travel information providing system that provides content selected by reliability estimation, which can increase satisfaction by easily providing high-quality travel information to potential travelers.

In addition, the present invention constructs and uses a reliability estimation model for estimating the reliability of content based on the importance of keywords extracted from travel-related content, the rating and recommendation of the creator, and thus the quality of the content itself and the quality of the content creator Another object of the present invention is to provide a travel information providing system that provides content selected by reliability estimation, which can provide more reliable information compared to other search engines or social media in consideration of all of the above.

A travel information providing system that provides content selected by reliability estimation according to a feature of the present invention for achieving the above object,

As a travel information providing system,

an information collection module for collecting travel-related content including travel course and accommodation reviews;

a model configuration module for constructing a reliability estimation model for estimating the reliability of the content based on the extracted keyword's importance, the author's evaluation level, and the recommendation level by applying machine learning to the collected travel-related contents;

a content selection module for selecting travel-related content using the constructed reliability estimation model; and

It is characterized in that it includes a content providing module that provides the selected travel-related content to travelers.

Preferably, the model configuration module,

(1) extracting a keyword from the travel-related content collected by the information collection module, and constructing a chunk including the importance of the keyword;

(2) calculating the rating of the author who wrote the travel-related content

(3) calculating a recommendation degree of the configured chunk;

(4) clustering chunks according to the rating of the author, the importance included in the chunk, and the recommendation degree of the chunk; and

(5) The reliability estimation model may be configured by performing the step of configuring the reliability estimation model by setting the reliability for each cluster according to the clustering result.

More preferably, the step (1) is

(1-1) classifying the travel-related content collected by the information collection module by subject;

(1-2) extracting keywords from the travel-related content and calculating the importance of the extracted keywords; and

(1-3) may include the step of constructing a chunk including the importance information of the keyword.

More preferably, in step (2),

The rating of the author may be calculated using at least one selected from the group including the rating, level, popularity, participation, intelligence, reliability, and number of content creation of the author on the platform where the travel-related content is written. .

Even more preferably, in step (2),

The author's reliability may be set differently according to at least one of the author's gender, age, and occupation.

More preferably, in step (3),

Based on at least one selected from the group including the number of recommendations, the number of comments, the number of pros and cons, and the number of views for the travel-related content related to the chunk, the chunk is classified into recommended, neutral, and non-recommended, and the degree of recommendation can be set according to the classification have.

More preferably, in step (4),

The chunks are clustered using Hierarchical Agglomerative Clustering, and the number of clusters may be determined using Elbow Criterion.

Preferably, the content selection module comprises:

Travel-related content may be searched for the input search word, and content having a reliability greater than or equal to a preset criterion may be selected from among the searched travel-related content using the reliability estimation model.

According to the travel information providing system that provides content selected by reliability estimation proposed in the present invention, travel selected using a reliability estimation model constructed by applying machine learning to travel-related content including travel course and accommodation reviews By providing relevant content, you can easily provide high-quality travel information to potential travelers to increase their satisfaction.

In addition, according to the travel information providing system that provides content selected by reliability estimation proposed in the present invention, the reliability of content is estimated based on the importance of keywords extracted from travel-related content, and the rating and recommendation of the author. By constructing and using a reliability estimation model that

1 is a view showing an overall configuration including a travel information providing system for providing content selected by reliability estimation according to an embodiment of the present invention.
2 is a diagram illustrating a detailed configuration of a travel information providing system for providing content selected by reliability estimation according to an embodiment of the present invention;
3 is a diagram illustrating a detailed flow of a process processed in a model configuration module of a travel information providing system that provides content selected by reliability estimation according to an embodiment of the present invention.
4 is a diagram illustrating a detailed flow of step S100 processed by a model configuration module of a travel information providing system that provides content selected by reliability estimation according to an embodiment of the present invention.

Hereinafter, preferred embodiments will be described in detail so that those of ordinary skill in the art can easily practice the present invention with reference to the accompanying drawings. However, in describing a preferred embodiment of the present invention in detail, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. In addition, the same reference numerals are used throughout the drawings for parts having similar functions and functions.

In addition, throughout the specification, when a part is 'connected' with another part, it is not only 'directly connected' but also 'indirectly connected' with another element interposed therebetween. include In addition, "including" a certain component means that other components may be further included, rather than excluding other components, unless otherwise stated.

1 is a diagram illustrating an overall configuration including a travel information providing system 10 that provides content selected by reliability estimation according to an embodiment of the present invention. As shown in FIG. 1 , the travel information providing system 10 for providing content selected by reliability estimation according to an embodiment of the present invention provides various platforms such as a web/app on which travel-related content is posted and shared. Through this, a plurality of travel-related contents may be collected, analyzed and selected, and provided to the user terminal 20 . The user terminal 20 accesses a website or application operated by the travel information providing system 10, requests a search by inputting search conditions, or selects a desired category, and selects a high-quality travel related according to reliability estimation. content can be provided.

Here, the user terminal 20 is a smartphone, a tablet PC (personal computer), a mobile phone, a video phone, an e-book reader, a desktop PC, a laptop PC, a netbook computer, It may include at least one of a workstation, a server, a personal digital assistant (PDA), a media box, a game console, an electronic dictionary, or a wearable device, and the wearable device is an accessory type (eg, Watches, rings, bracelets, anklets, necklaces, glasses, contact lenses), or head mounted devices (HMDs), integrated with textiles or clothing (e.g. electronic garments), body-worn (e.g. skin pads) pad or tattoo), or at least one of an implantable circuit.

However, the user terminal 20 of the present invention is not limited to the form of the terminal as listed, and the travel information providing system 10 is connected through a network to receive and output selected travel-related content using a reliability estimation model. If possible, the role of the user terminal 20 of the present invention can be freely performed regardless of the specific terminal type. Also, the user terminal 20 may be a combination of two or more of the various devices described above.

On the other hand, the network is a wired network such as a local area network (LAN), a wide area network (WAN) or a value added network (VAN), or a mobile communication network (mobile radio communication network), a satellite communication network, It may be implemented in all types of wireless networks such as Bluetooth, Wireless Broadband Internet (Wibro), High Speed Downlink Packet Access (HSDPA), Long Term Evolution (LTE), and 5th Generation Mobile Telecommunication (5G).

2 is a diagram illustrating a detailed configuration of a travel information providing system 10 that provides content selected by reliability estimation according to an embodiment of the present invention. As shown in FIG. 2 , the travel information providing system 10 for providing content selected by reliability estimation according to an embodiment of the present invention includes an information collection module 100 , a model construction module 200 , and content It may be configured to include a selection module 300 and a content providing module 400 .

The information collection module 100 may collect travel-related content including travel course and accommodation reviews. More specifically, the information collection module 100 may collect travel-related content on the web through a crawler, and may collect various types of travel-related content including text, images, and videos. In addition, the travel-related content may be a variety of content regardless of the specific type, as long as it is travel-related, such as a travel course and accommodation review, as well as a restaurant review, car rental review, airline review, tourist destination review, and travel tip.

The model configuration module 200 applies machine learning to the collected travel-related content to construct a reliability estimation model for estimating the reliability of the content based on the importance of the extracted keyword, the author's evaluation degree, and the recommendation degree. have. Hereinafter, with reference to FIG. 3 , a detailed flow of the process in which the model construction module 200 constructs a reliability estimation model will be described in detail.

3 is a diagram illustrating a detailed flow of a process processed by the model configuration module 200 of the travel information providing system 10 for providing content selected by reliability estimation according to an embodiment of the present invention. As shown in FIG. 3 , the model configuration module 200 of the travel information providing system 10 that provides content selected by reliability estimation according to an embodiment of the present invention extracts keywords from travel-related content and Composing a chunk including the importance of the keyword (S100), calculating the rating of the author who wrote the travel-related content (S200), calculating the recommendation of the chunk (S300), the author's rating, the chunk The reliability estimation model may be configured by processing the steps of clustering chunks ( S400 ) and configuring the reliability estimation model by setting the reliability for each cluster ( S500 ) according to the importance included in and the recommendation degree of the chunks.

In step S100 , a keyword may be extracted from the travel-related content collected by the information collection module 100 , and a chunk including the importance of the keyword may be configured. Hereinafter, the detailed flow of step S100 will be described in detail with reference to FIG. 4 .

4 is a diagram illustrating a detailed flow of step S100 processed by the model configuration module 200 of the travel information providing system 10 for providing content selected by reliability estimation according to an embodiment of the present invention. As shown in FIG. 4 , step S100 of the travel information providing system 10 for providing content selected by reliability estimation according to an embodiment of the present invention is a step of classifying travel-related content by subject (S110). , extracting a keyword from travel-related content, calculating the importance of the extracted keyword (S120), and configuring a chunk including the importance information of the keyword (S130).

In step S110, the travel-related content collected by the information collection module 100 may be classified by subject. For example, you can first classify by travel location, then classify by various topics, such as classifying by content target such as travel courses and accommodation reviews, and classifying by content characteristics such as things to do, tourist attractions, and restaurants. have.

In step S120, a keyword may be extracted from the travel-related content, and the importance of the extracted keyword may be calculated. In step S120, the keywords reflect the core contents of travel-related content, and keywords may be extracted through text mining. More specifically, in step S120, keywords are extracted using TF (Term Frequency), IDF (Inverse Document Frequency), TF-IDF (Term Frequency-Inverse Document Frequency), abstract summarization using deep learning, etc. , it is possible to calculate the importance of each keyword. In some embodiments, key keywords may be preset for each subject, and the importance of each key keyword may be calculated using text mining such as TF, IDF, and RF-IDF.

In step S130, a chunk including the importance information of the keyword may be configured. That is, in step S130, the importance of the keyword calculated in step S120 and the keyword may be configured in chunks. In this case, by configuring one chunk for each travel-related content, information on each travel-related content may be expressed in one chunk. In addition, since importance information of several keywords is included in a chunk, the importance information of several keywords may be aggregated to calculate the importance of a chunk.

Additionally, the model configuration module 200 may preprocess the chunk configured in step S130 and separate it into morpheme units. More specifically, the chunk may be preprocessed through vectorization, normalization, missing value processing, and feature extraction, and the chunk may be separated into morpheme units.

In step S200, it is possible to calculate the evaluation degree of the creator who created the travel-related content. That is, in step S200, the content source that created the travel-related content may be evaluated. If the evaluation degree of the author calculated in step S200 is high, it may be determined that the author is highly likely to create high-quality content.

More specifically, in step S200, the rating of the author using at least one selected from the group including the rating, level, popularity, participation, intelligence, reliability, and the number of content creation on the platform where travel-related content is written can be calculated. In this case, in step S200, the author's reliability may be set differently according to at least one of the author's gender, age, and occupation. That is, the reliability in step S200 means the reliability of the source for how reliable the creator of the content is, and may be a different concept from the reliability of the travel-related content.

In step S300, a recommendation degree of the configured chunk may be calculated. That is, in step S300 , the recommendation degree of the travel-related content itself is evaluated, but the recommendation degree of the chunk is calculated based on the recommendation degree of the travel-related content, so that learning and processing can be performed in chunks.

More specifically, in step S300, based on at least one selected from the group including the number of recommendations, the number of comments, the number of pros and cons, and the number of views for travel-related content related to the chunk, the chunk is classified into recommended, neutral, and non-recommended, and the classification is performed. Recommendations can be set accordingly. In this case, the classification criteria and the recommendation degree according to the classification may be preset.

In step S400, the chunks may be clustered according to the rating of the author, the importance included in the chunk, and the recommendation degree of the chunk. That is, in step S400, the importance of the keyword calculated in step S100 to compose the chunk, the evaluation degree of the creator calculated in step S200, the recommendation degree of the chunk calculated in step S300, etc. of machine learning for unsupervised learning By input to a kind of clustering algorithm, chunks can be clustered. At this time, since importance information of several keywords is included in the chunk, the importance of the chunks can be used for clustering by using the sum and average of the importance of several keywords included in the chunk as the importance of the chunk.

More specifically, in step S400, chunks are clustered using Hierarchical Agglomerative Clustering, and the number of clusters may be determined using Elbow Criterion.

Hierarchical clustering is a method of constructing clusters hierarchically by connecting data, and there is no need to assume the number of clusters in the first place. In hierarchical clustering, there are Agglomerative and Divisive. Agglomerative hierarchical clustering assumes that each individual data in the given data is an independent cluster, and merges them by an algorithm to form a higher-level cluster. It is a method of constructing a cluster with all data as members by constructing the cluster and merging the upper-level cluster again by an algorithm. In the present invention, while constructing clusters using cohesive hierarchical clustering, elbow evaluation may be used to determine the optimal number of clusters. Accordingly, in step S400, all chunks may be clustered by the number of clusters determined by the elbow evaluation.

In step S500, a reliability estimation model may be configured by setting the reliability for each cluster according to the clustering result. That is, in step S500, the reliability for each cluster classified according to the number of clusters determined by the elbow evaluation in step S400 may be set. In this case, if the number of clusters is three, the reliability of each cluster may be set to high, medium, or low reliability.

The content selection module 300 may select travel-related content by using the constructed reliability estimation model. More specifically, the content selection module 300 may search for travel-related content with respect to the input search word, and select content having a reliability equal to or greater than a preset criterion among the searched travel-related content using a reliability estimation model. Here, the content selection module 300 may perform a search by receiving a search term that the traveler wants to search from the user terminal 20 .

In addition, the content selection module 300 inputs the searched travel-related content into the model configuration module 200, and through the morpheme analysis of the chunk configured in step S100, the analyzed morpheme pattern is collected by the information collection module 100, If it is a new pattern that does not match the analyzed existing pattern, the model configuration module 200 may update the reliability estimation model by re-performing clustering using the searched travel-related content as training data. That is, by continuously adding new travel-related content as training data, it is possible to improve the accuracy of the reliability estimation model.

On the other hand, the content selection module 300, for example, when the reliability is divided into high, medium, and low, may select the travel-related content corresponding to the prize, and if the content corresponding to the prize is less than a predetermined number, You can select the content that corresponds to the top and the middle. In this way, the content selection module 300 may provide a sufficient amount of content by adjusting a preset criterion according to the number of selected content.

The content providing module 400 may provide the selected travel related content to the traveler. The content providing module 400 may provide the selected travel-related content to the user terminal 20 through a network as shown in FIG. 1 . In this case, the travel-related content may be provided by being classified by subject, and if the content has the same reliability, the content may be arranged and provided in the order of the latest creation date.

As described above, according to the travel information providing system 10 that provides content selected by reliability estimation proposed in the present invention, machine learning is applied to travel-related content including travel course and accommodation reviews. By providing selected travel-related content using a reliability estimation model, it is possible to easily provide high-quality travel information to potential travelers to increase their satisfaction.

In addition, according to the travel information providing system 10 that provides content selected by reliability estimation proposed in the present invention, the content of the content is based on the importance of keywords extracted from travel-related content, and the rating and recommendation of the author. By constructing and using a reliability estimation model for estimating reliability, it is possible to provide reliable information compared to other search engines or social media by considering both the quality of the content itself and the quality of the content creator.

Various modifications and applications of the present invention described above are possible by those skilled in the art to which the present invention pertains, and the scope of the technical idea according to the present invention should be defined by the following claims.

10: Travel information provision system
20: user terminal
100: information collection module
200: model configuration module
300: content selection module
400: content providing module
S100: extracting keywords from travel-related content and composing chunks including the importance of keywords
S110: Classifying travel-related content by topic
S120: Step of extracting keywords from travel-related content and calculating the importance of the extracted keywords
S130: Step of constructing a chunk including the importance information of the keyword
S200: Step of calculating the rating of the author who wrote the travel-related content
S300: Calculating the recommendation degree of the chunk
S400: Clustering the chunks according to the rating of the author, the importance contained in the chunk, and the recommendation of the chunk
S500: A step of configuring a reliability estimation model by setting the reliability for each cluster

Claims (8)

As a travel information providing system (10),
an information collection module 100 for collecting travel-related content including travel course and accommodation reviews;
a model configuration module 200 for constructing a reliability estimation model for estimating the reliability of the content based on the extracted keyword's importance, the author's evaluation degree, and the recommendation degree by applying machine learning to the collected travel-related content;
a content selection module 300 for selecting travel-related content by using the constructed reliability estimation model; and
Includes a content providing module 400 that provides the selected travel-related content to travelers,
The model configuration module 200,
(1) extracting a keyword from the travel-related content collected by the information collection module 100, and constructing a chunk including the importance of the keyword;
(2) calculating an evaluation degree of the author who wrote the travel-related content;
(3) calculating a recommendation degree of the configured chunk;
(4) clustering chunks according to the rating of the author, the importance included in the chunk, and the recommendation degree of the chunk; and
(5) configuring the reliability estimation model by configuring the reliability estimation model by setting the reliability for each cluster according to the clustering result;
The step (1) is,
(1-1) classifying the travel-related content collected by the information collection module 100 by subject;
(1-2) Extracting preset key keywords for each subject from the travel-related content, and abstracting them using TF (Term Frequency), IDF (Inverse Document Frequency), TF-IDF (Term Frequency-Inverse Document Frequency) and deep learning calculating the importance of the extracted keyword using at least one of abstract summarization; and
(1-3) Configure a chunk including the importance information of the keyword, but configure one chunk for each travel-related content including the importance information of a plurality of keywords extracted from the travel-related content Including the step of expressing the information in one chunk,
In step (3),
Based on the recommendation for the travel-related content, the chunk recommendation is calculated to enable learning and processing in chunk units, and includes the number of recommendations, comments, pros and cons, and views for travel-related content related to the chunk. Classifying the chunks into recommended, neutral, and non-recommended based on at least one selected from the group, and setting a recommendation level according to the classification,
The content selection module 300,
Searching for travel-related content with respect to the input search word, and selecting content with a reliability higher than or equal to a preset criterion from among the searched travel-related content using the reliability estimation model,
The content selection module 300 inputs the searched travel-related content into the model construction module 200, and through morphological analysis of the chunk constructed in step (1), the analyzed morpheme pattern is generated by the information collection module 100. If it is a new pattern that does not match the existing pattern collected and analyzed by , a travel information providing system 10 that provides content selected by reliability estimation.
delete delete The method of claim 1, wherein in step (2),
Computing the rating of the author using at least one selected from the group including the rating, level, popularity, participation, intelligence, reliability, and number of content creation of the author on the platform where the travel-related content is written A travel information providing system 10 that provides content selected by reliability estimation.
The method of claim 4, wherein in step (2),
A travel information providing system (10) for providing content selected by reliability estimation, characterized in that the author's reliability is set differently according to at least one of the author's gender, age, and occupation.
delete The method of claim 1, wherein in step (4),
Travel information providing content selected by reliability estimation, characterized in that the chunks are clustered using Hierarchical Agglomerative Clustering and the number of clusters is determined using Elbow Criterion system (10). delete

Images