JP6373197B2 - Comment classification program, server, and method for extracting map route related comments from multiple comments - Google Patents

Description

  The present invention relates to a technique for classifying comments posted from an unspecified number of third parties.

  On the Internet, a blog (Web log) server or a mini blog (for example, twitter (registered trademark) or facebook (registered trademark)) server is connected. Such a blog server receives comments from an unspecified number of third parties and publishes them to other third parties. These comments are written for the purpose of publishing various topics that occur in current affairs. For example, topics that are closely related to map routes, such as train suspension and delays and traffic congestion, are often described.

  Conventionally, there is a technique of performing topic classification only on comments (sentences) to which position information is added and extracting topics and terms related to the area (see, for example, Non-Patent Document 1). According to this technology, topics related to the area are extracted from a large number of comments with position information, and various social phenomena can be discovered.

J. Eisenstein, et.al. A Latent Variable Model for Geographic Lexical Variation, EMNLP2010, [online], [searched on January 21, 2015], Internet <URL: http://www.cs.cmu.edu/ ~ nasmith / papers / eisenstein + oconnor + smith + xing.emnlp10.pdf> DM Blei, AY Ng, and MI Jordan. Journal of Machine Learning Research, 3: 993-1022, 2003, [online], [Search January 21, 2015], Internet <URL: http: / /machinelearning.wustl.edu/mlpapers/paper_files/BleiNJ03.pdf>

  However, the number of comments with position information relative to the total number of comments is less than 1%. If a topic related to the region is extracted from comments sent from only a small number of contributors, the information itself may be unevenly distributed. Actually, according to the technique described in Non-Patent Document 1, even if a slang term unique to a region and the name of a local baseball team can be extracted, it is difficult to extract a topic that is currently occurring in that region. It is also difficult to extract topics such as traffic paralysis and traffic congestion that occur in map routes, such as transportation and roads. In particular, only a very small number of comments with position information is not useful because the accuracy of information reliability is low.

  In general, comments posted from the same map route are considered to have similar topics regardless of position, with or without position information. On the other hand, the inventors of this application can distinguish the topic regarding the map route according to the similarity of the topic in the comment that occurs at present. I thought.

  Therefore, an object of the present invention is to provide a comment classification program, a server, and a method capable of extracting a map route related comment from a large number of comments in which comments with position information and comments without position information are mixed.

According to the present invention, a plurality of comments and positional information without comment and comment-located are mixed, a comment classification program that causes a computer to function so as to extract a map routing related comments,
Route position storage means for storing a plurality of positions on the route for each map route;
A topic classification means for classifying a plurality of comments into a plurality of topic groups based on a distribution of word frequencies, and assigning a topic number to each comment;
Comment classification means with position information for classifying a plurality of comments with position information for each map route of the route position storage means included in a predetermined range from the position information given to the comment;
For each map path, in commented positional information of multiple, and maps the route topic number extracting means for extracting a predetermined criterion highest topic number based on localization of to the map route as a map route topic number,
The computer is caused to function as a map route comment classification means for classifying comments with location information and comments without location information as comments related to the map route based on the map route topic number of each map route.

According to another embodiment of the comment classification program of the present invention,
The topic classification means causes the computer to function so as to classify each comment into any one topic group using an LDA (Latent Dirichlet Allocation) algorithm that calculates the probability (topic ratio) belonging to each topic group. Is also preferable.

According to another embodiment of the comment classification program of the present invention,
The predetermined criterion based on the uneven distribution degree in the map route topic number extraction means is the Akaike information amount criterion, and one of the following criteria is used (first criterion: number of comments).
n11 = number of comments for map route i included in topic number j n12 = number of comments for map route i included other than topic number j n21 = number of comments other than map route i included in topic number j n22 = other than topic number j Number of comments other than map route i included in (second criterion: total evaluation value)
n11 = total evaluation value for map path i of comments included in topic number j n12 = total evaluation value for map path i of comments included other than topic number j n21 = map path i of comments included in topic number j Total of evaluation values for other than n22 = It is also preferable to make the computer function like a total of evaluation values for other than the map route i of comments included other than the topic number j.

According to another embodiment of the comment classification program of the present invention,
The map route topic number extraction means calculates the centroid of the position coordinates of a plurality of comments with position information having map route topic numbers extracted for each map route, and calculates the average value of the distance from the centroid to the position coordinates of each comment. When the calculated average value is equal to or less than a predetermined threshold value, it is also preferable that the computer function so that the map route topic number is not extracted because it is not a map route.

According to another embodiment of the comment classification program of the present invention,
The topic classification means preferably analyzes the predicate term structure of the comment with position information and the comment without position information, and causes the computer to function so as to remove the comment including the hearing estimation expression.

According to another embodiment of the comment classification program of the present invention,
It is also preferable to make the computer function so that the comment is posted to a mini blog (mini Web log) server by an unspecified number of third parties.

According to the present invention, and collecting a plurality of comments and positional information without comment and comment-located are mixed, a comment classification server for extracting the map route associated comments,
Route position storage means for storing a plurality of positions on the route for each map route;
A topic classification means for classifying a plurality of comments into a plurality of topic groups based on a distribution of word frequencies, and assigning a topic number to each comment;
Comment classification means with position information for classifying a plurality of comments with position information for each map route of the route position storage means included in a predetermined range from the position information given to the comment;
For each map path, in commented positional information of multiple, and maps the route topic number extracting means for extracting a predetermined criterion highest topic number based on localization of to the map route as a map route topic number,
It has map route comment classification means for classifying comments with position information and comments without position information as comments related to the map route based on the map route topic number of each map route.

According to the present invention, and collecting a plurality of comments and positional information without comment and comment-located are mixed, a comment classification method of the device for extracting the map route associated comments,
The device
Each map route has a route position storage unit that stores a plurality of positions on the route,
A first step of classifying a plurality of comments into a plurality of topic groups based on the distribution of word frequencies, and assigning a topic number to each comment;
A second step of classifying a plurality of comments with position information for each map route in the route position storage unit included in a predetermined range from the position information given to the comment;
For each map path, in commented positional information of multiple, a third step of extracting a predetermined criterion highest topic number based on localization of to the map route as a map route topic number,
And a fourth step of classifying comments with position information and comments without position information as comments related to the map path based on the map path topic number of each map path.

  According to the comment classification program, server, and method of the present invention, map route-related comments can be extracted from a large number of comments in which comments with position information and comments without position information are mixed.

It is a system configuration diagram in the present invention. It is a functional block diagram of the comment classification | category server in this invention. It is explanatory drawing showing the example of the comment accumulate | stored in the comment storage part. It is explanatory drawing showing the classification | category of the comment in a topic classification | category part. It is explanatory drawing represented as a comment with position information and a map route topic number. It is explanatory drawing using the stochastic information criterion. It is explanatory drawing which specifies the comment classified by the map route topic number. It is a sequence diagram in the present invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a system configuration diagram according to the present invention.

According to FIG. 1, a comment classification server 1 according to the present invention is connected to the Internet. The comment classification server 1 collects a large number of comments from the blog server 2 that publishes comments posted by an unspecified number of third parties. The comment classification server 1 can classify map route related comments from a large number of comments in which comments with position information (sentences) and comments without position information (sentences) are mixed.
The blog server 2 is a mini blog server such as twitter (registered trademark) or facebook (registered trademark).
The terminal 3 is owned by an unspecified number of third parties, and can freely post comments to the miniblog server 2.
The terminal 4 is operated by the viewer, can access the comment classification server 1 via the access network and the Internet, and can receive a desired map route related comment.

A large number of comments are roughly classified into the following three patterns even if they are topics (topics) based on the same position.
(Topic 1) Comments spread throughout the region (for example, topics not specified in the region)
(Topic 2) Comments spread along the map route (for example, topics based on transportation)
(Topic 3) A group of comments in one place (for example, topics of people gathered at a local event)
Here, the comment classification server 1 of the present invention classifies the comment group that has spread along the (topic 2) map route.

  FIG. 2 is a functional configuration diagram of the comment classification server in the present invention.

  As shown in FIG. 2, the comment classification server 1 includes a communication interface unit 10, a comment collection unit 101, and a comment storage unit 102.

[Comment Collection Unit 101]
The comment collection unit 101 receives comments posted from an unspecified number of third parties from the blog server 2 via the network, and outputs the comments to the comment storage unit 102. Note that the comment collection unit 101 may collect only comments including a plurality of keywords specified in advance corresponding to a desired application.

[Comment storage unit 102]
FIG. 3 is an explanatory diagram illustrating an example of comments accumulated in the comment accumulation unit.

  The comment accumulation unit 102 accumulates a large amount of comments collected by the comment collection unit 101. The “comment” is, for example, a Japanese “tweet” (maximum number of characters: 140 characters) transmitted by twitter (registered trademark). The comment includes at least a user id (from_user_id), a tweet ID (id_str), a transmission time (created_at), and a tweet (texts). Here, the comment may be “with position information” or “without position information”. The position information is generally latitude / longitude information acquired by GPS (Global Positioning System) in the comment poster terminal. Note that the position information is not limited to this, and may be position information based on positioning by a plurality of base stations, or position information associated with an access point.

  The comment storage unit 102 preferably divides a large number of comments into comment groups for each predetermined time range according to the transmission time. For example, it may be every hour. Thereby, it is possible to know the change in the map route related comment every hour.

  According to FIG. 2, the comment classification server 1 further includes a morphological analysis unit 110, a topic classification unit 111, a comment classification unit with position information 112, a route position storage unit 103, a map route topic number extraction unit 113, And a map route comment classification unit 114. These functional components are realized by executing a program that causes a computer mounted on the server to function. Further, the processing flow of these functional components can be understood as a comment classification method of the apparatus.

[Morphological analyzer 110]
The morpheme analysis unit 110 classifies a large number of comments in the comment storage unit 102 into morphemes by morpheme analysis. “Morphological analysis” refers to a technique in which sentences are divided into meaningful words and the part of speech and contents are discriminated using a dictionary. The “morpheme” means the smallest unit having meaning among the elements of the sentence. The classified morphemes are output to the topic classification unit 111.

[Topic classification unit 111]
The topic classification unit 111 classifies a large number of comments into a plurality of topic groups based on the distribution of word appearance frequencies, and assigns a topic number to each comment.

<LDA analysis of topic classification unit 111>
The topic classification unit 111 classifies each comment into any one topic group using an LDA (Latent Dirichlet Allocation) algorithm that calculates a probability (topic ratio) belonging to each topic group. Topic classification is a probabilistic model of a document generation process expressed in bag-of-words in order to analyze comments (see, for example, Non-Patent Document 2).

  LDA introduces a probabilistic framework based on fluctuations in a feature vector of a word to a technique (LSI (latent Semantic Indexin)) for dimensional compression of a word document matrix. The compressed set of dimensions is called a topic.

The topic classification unit 111 executes processing in the following steps.
(S1) From a large number of comments, the appearance frequency (number of appearances) for each word is input to the LDA process. And the appearance frequency of each word is counted for every comment.
(S2) Next, the word distribution for each topic and the topic ratio for each comment are acquired. This topic ratio classifies the topic group to which the comment belongs. And the appearance frequency of each word contained in all the comments is counted for every topic group.
(S3) Finally, the words belonging to each topic group are counted for each comment. Then, the comments are classified into topic groups with high count values.

  FIG. 4 is an explanatory diagram showing the classification of comments in the topic classification unit.

According to FIG. 4, the topic classification unit 111 classifies the large number of comments in FIG. 3 into three topic groups as a result. Each comment is given a classified topic number.
Then, the comments with position information are output to the comment classification section with position information 112 and the map route comment classification section 114.
Further, the comment without position information is output to the map route comment classification unit 114.

<Predicate term structure analysis of topic classifier 111>
It is also preferable that the topic classifying unit 111 further analyzes the predicate term structure of the comment with position information and the comment without position information, and removes the comment including the hearing estimation expression.

  “Predicate term structure” refers to a noun phrase or the like that becomes a “term” to a predicate in a sentence. By using the predicate term structure, it is possible to grasp the skeleton of the meaning of the sentence. The predicate term structure is composed of “object” for “predicate” and its case. As the predicate term structure analysis, for example, a predicate term structure analyzer such as Syncha which is free software can be used.

Instead of expressing a document as a bug of words (BOW), topic classification is performed using a predicate term structure. The following predicate term structure is acquired from each example sentence.
For example, a comment is analyzed as follows.
Comment 1 “Helicopters are gathering”
-> Predicate “Gather (situation)”, Ga ’s “Helicopter”
Comment 2 “Helicopters are gathering”
-> Predicate “Gather (situation)”, Ga ’s “Helicopter”
Comment 3 “Helicopters are gathering”
-> Predicate “Gather (hearing)”, Ga ’s “Helicopter”
Comment 4 “Helicopters seem to be gathering”
-> Predicate “Gather (estimate)”, Ga ’s “Helicopter”

  According to the above-described example, even if the same predicate is “gather”, it is possible to distinguish between “gather (situation)” of comments 1 and 2 and “gather (hearing / estimation)” of comments 3 and 4. . If BOW, which is an existing topic classification, is used, the feature quantities of these comments are the same and cannot be analyzed separately. That is, according to the present invention, the comments 1 and 2 and the comments 3 and 4 can be distinguished and classified into topics, and different topic numbers are assigned.

  According to the predicate term structure analysis, the comments 1 and 2 can be analyzed as situation expressions posted from the map path of the site. Moreover, the comments 3 and 4 can be analyzed to be hearing estimation expressions posted from positions other than the map route of the site. In this case, the comments 3 and 4 including the hearing estimation expression are removed. As a result, only comments posted from the map route of the site (with or without position information) can be associated on the map.

  FIG. 5 is an explanatory diagram illustrating a comment with position information and a map route topic number.

[Route location storage unit 103]
The route position storage unit 103 stores a plurality of positions on the route for each map route. According to FIG. 5, for each route, a plurality of pieces of position information that are continuous on the route are recorded.

[Comment classification unit 112 with position information]
The comment classifier 112 with position information classifies a plurality of comments with position information for each predetermined map route to which the position information belongs.

  According to FIG. 5, the position information of the comments with position information is plotted on the map. The map is divided in advance for each predetermined map route. As a result, the comments with position information included in the range are classified for each predetermined map route.

[Map route topic number extraction unit 113]
The map route topic number extraction unit 113 extracts, for each predetermined map route, the largest topic number as a map route topic number from among a plurality of comments with position information using a predetermined criterion based on the uneven distribution degree. The predetermined criterion based on the uneven distribution degree may be, for example, Akaike's Information Criterion (AIC).

  Here, the topic numbers of the comments with position information assigned to the map routes C1, C2,... Are t1, t2,. In the following, a method of calculating an index Info (C, t) for determining whether the topic number t is useful for determining the map route C will be described below.

(S1) From the set U of comments included in the map route, the following four types of frequencies are obtained according to any one of the following criteria.
(First criterion: number of comments)
n11 = number of comments for map route i included in topic number j n12 = number of comments for map route i included other than topic number j n21 = number of comments other than map route i included in topic number j n22 = other than topic number j The number of comments other than the map route i included in the number of comments as in the first criterion is based on the conventional AIC and is deterministically determined.
(Second criterion: total evaluation value)
n11 = total evaluation value for map path i of comments included in topic number j n12 = total evaluation value for map path i of comments included other than topic number j n21 = map path i of comments included in topic number j N22 = total evaluation value for comments other than topic number j other than map route i other than map number i Based on the evaluation value corresponding to the map route as in the second criterion, probability theory It is discriminate | determined automatically.

For example, it is assumed that the topic number j includes comments a, b, and c. At this time, it is assumed that the posting positions of the comments a and c are very close to the map route i, and the posting position of the comment b is away from the map route i. Note that comment a is very close to map route i and has an evaluation value of 0.9, and comment c is close to map route i and has an evaluation value of 0.8. The comment c is away from the map route i and has an evaluation value of 0.1. Here, n11 of the first criterion and the second criterion is calculated as follows, for example.
First criterion: n11 = 1 + 0 + 1 = 2
Second criterion: n11 = 0.9 + 0.1 + 0.8 = 1.8

  FIG. 6 is an explanatory diagram using a stochastic information criterion.

According to FIG. 6, the comment with the location information of the topic number j is posted from the suburbs of the Tojo Line and Kawagoe Line. Also, position coordinate points are arranged on each map route. Here, the topic number j can be evaluated as, for example, a probability 0.3 corresponding to the map route 1 (Kawagoe Line) and a probability 0.7 corresponding to the map route 2 (Tojo Line). For example, for topic number j, 70% is determined to be based on map route 2 and 30% is determined to be based on map route 1.
Li: The shortest distance between the position coordinates of a comment and the position coordinates of the map route i Vi = 1 / Li: Evaluation value of whether a comment belongs to the map route i That is, the position information always has an error. This is because it cannot be determined that it is included and is not necessarily close to one map route. There is also a possibility that the topic number j refers to a plurality of map routes.

(S2) Next, with respect to n11, n12, n21, and n22, the value MLL_IM (C, t) for the independent model and the value MLL_DM (C, t) for the dependent model are calculated using the Akaike information criterion. This calculates an unreasonable ratio for each set of map route and topic number.
MLL_IM (C, t) = (n11 + n12) log (n11 + n12)
+ (N11 + n21) log (n11 + n21)
+ (N21 + n22) log (n21 + n22)
+ (N12 + n22) log (n12 + n22) -2N log N
MLL_DM (C, t) = n11 log n11 + n12 log n12 + n21 log n21 + n22 log n22−N log N
However, N = n11 + n12 + n21 + n22

(S3) The following Info (C, t) is calculated from the aforementioned MLL_IM (C, t) and MLL_DM (C, t).
AIC_IM (C, t) = -2 × MLL_IM (C, t) + 2 × 2
AIC_DM (C, t) = -2 × MLL_DM (C, t) + 2 × 3
Info (C, t) = AIC_IM (t, C) − AIC_DM (C, t)

  Info (C, t) calculated above represents an unreasonable proportion in which the topic number t appears with a bias toward the map route C. Info (C, t) has a higher Info (C, t) value as the topic number t is useful for determining the map route C according to the Akaike information criterion. According to the present invention, for each map route Ci, m topic numbers t can be extracted in descending order of Info (C, t).

<Removal of Concentrated Topic at Map Route Topic Number Extraction Unit 113>
It is also preferable that the map route topic number extraction unit 113 does not extract topic numbers that concentrate on one position coordinate as a map route topic number because it is not a map route (for example, not a transportation facility). Specifically, the following steps are executed.
(S1) The center of gravity of position coordinates of a plurality of comments with position information having map path topic numbers extracted for each map path is calculated.
(S2) The average value of the distance to the position coordinates of each comment is calculated from the centroid.
(S3) If the average value is equal to or less than the predetermined threshold value, the map route topic number is not extracted because it is not a map route.

  For example, when a fire breaks out near a certain route (map route), a large number of twitter comments posted during that time are collected. At this time, even a comment without position information can be associated with the actual position of the fire site. However, topic numbers are concentrated at one position coordinate. According to the present invention, in this case, the topic number is not extracted because it is not based on the map route.

[Map route comment classification unit 114]
Based on the map route topic number of each map route, the map route comment classification unit 114 classifies the comment with position information and the comment without position information as a map route related comment.

  FIG. 7 is an explanatory diagram clearly showing comments classified by map route topic numbers.

  According to FIG. 7, for example, it is assumed that the number of topic numbers 1 is the largest among the comments with position information located in the map route a. In this case, it can be estimated that the comment of the topic number 1 is posted from within the map route a. Thereby, the comment with the location information of topic number 1 and the comment without the location information can be classified as a map route related comment on the map route a.

  For example, when a vehicle failure occurs on a certain route (map route), it is assumed that a large number of twitter comments posted during that time period are collected. At this time, out of about 20,000 comments, there are only about 50 comments with location information. By applying the present invention, topic classification of LDA is executed for about 20,000 comments, and the same comment as the topic number of about 50 comments with position information can be associated with the position information. As a result, even a comment without position information can identify a map path that is actually paralyzed and view the comment content. You can know the cause of the traffic paralysis from the comments.

  FIG. 8 is a sequence diagram in the present invention.

According to FIG. 8, a comment classification server process (a comment classification method of the apparatus) that collects a large number of comments in which comments (sentences) with position information and comments (sentences) without position information are mixed and extracts map route related comments. ).
(S111) First, the comment classification server 1 classifies a large number of comments into a plurality of topic groups based on the distribution of word appearance frequencies, and assigns topic numbers to the comments (topics in FIG. 2 described above). Classification unit 111).
(S112) Next, the comment classification server 1 classifies a plurality of comments with position information for each map route in the route position storage means included in a predetermined range from the position information given to the comment (the above-described figure). 2 (see comment classification unit 112 with position information).
(S113) Next, for each map route, the comment classification server 1 uses a predetermined criterion based on the degree of uneven distribution, and extracts the largest topic number as a map route topic number among a plurality of comments with position information ( (See the map route topic number extraction unit 113 in FIG. 2 described above).
(S114) Based on the map route topic number of each map route, the comment classification server 1 classifies the comment with position information and the comment without position information as the map route related comment (the map route in FIG. 2 described above). (See comment classification unit 114).
Then, the terminal 4 operated by the viewer accesses the comment classification server 1 via the access network and the Internet. The terminal 4 can receive the map route related comment of the map route from the comment classification server 1 by transmitting the map route as a query to the comment classification server 1.

  As described above in detail, according to the comment classification program, server, and method of the present invention, it is possible to extract map route related comments from a large number of comments in which comments with position information and comments without position information are mixed. it can. In particular, according to the present invention, even if a comment does not necessarily include a word specifying a position such as a route name or a station name, a topic that is unevenly distributed on a map route can be detected based on the topic classification.

  Various changes, modifications, and omissions of the above-described various embodiments of the present invention can be easily made by those skilled in the art. The above description is merely an example, and is not intended to be restrictive. The invention is limited only as defined in the following claims and the equivalents thereto.

DESCRIPTION OF SYMBOLS 1 Comment classification | category server 10 Communication interface part 101 Comment collection part 102 Comment storage part 103 Path | route position memory | storage part 110 Morphological analysis part 111 Topic classification | category part 112 Comment classification | category part with location information 113 Map route topic number extraction part 114 Map route comment classification part 2 Mini-blog server 3 Terminal for contributors 4 Terminal for viewers

Claims (8)

A plurality of comments with position information comments and the positional information without comment are mixed, a comment classification program that causes a computer to function so as to extract a map routing related comments,
Route position storage means for storing a plurality of positions on the route for each map route;
A topic classification means for classifying a plurality of comments into a plurality of topic groups based on a distribution of word frequencies, and assigning a topic number to each comment;
Comment classification means with position information for classifying a plurality of comments with position information for each map route of the route position storage means included in a predetermined range from the position information given to the comment;
For each map path, in commented positional information of multiple, and maps the route topic number extracting means for extracting a predetermined criterion highest topic number based on localization of to the map route as a map route topic number,
A comment classification program for causing a computer to function as a map route comment classification means for classifying comments with location information and comments without location information as comments related to the map route based on the map route topic number of each map route . The topic classification means causes the computer to function so as to classify each comment into any one topic group using an LDA (Latent Dirichlet Allocation) algorithm that calculates a probability (topic ratio) belonging to each topic group. The comment classification program according to claim 1, wherein: The predetermined criterion based on the degree of uneven distribution in the map route topic number extraction means is the Akaike information amount criterion, and uses one of the following criteria (first criterion: number of comments)
n11 = number of comments for map route i included in topic number j n12 = number of comments for map route i included other than topic number j n21 = number of comments other than map route i included in topic number j n22 = other than topic number j Number of comments other than map route i included in (second criterion: total evaluation value)
n11 = total evaluation value for map path i of comments included in topic number j n12 = total evaluation value for map path i of comments included other than topic number j n21 = map path i of comments included in topic number j 3. The comment classification program according to claim 1 or 2, wherein the computer is caused to function as a sum of evaluation values for other than the map path i of comments included other than the topic number j. The map route topic number extraction means calculates the centroid of the position coordinates of a plurality of comments with position information having the map route topic number extracted for each map route, and averages the distance from the centroid to the position coordinates of each comment The computer is caused to function so as not to extract a map route topic number, assuming that the map route is not a map route when the value is calculated and the average value is a predetermined threshold value or less. Comment classification program described in. 5. The topic classifying unit analyzes a predicate term structure of a comment with position information and a comment without position information, and causes the computer to function to remove a comment including a hearing estimation expression. The comment classification program according to item 1. The computer according to any one of claims 1 to 5, wherein the computer functions so that the comment is posted to a mini blog (mini Web log) server by a plurality of unspecified third parties. The comment classification program described. Location information with comments and the position information without comment to collect multiple comments are mixed, a comment classification server to extract the map route-related comments,
Route position storage means for storing a plurality of positions on the route for each map route;
A topic classification means for classifying a plurality of comments into a plurality of topic groups based on a distribution of word frequencies, and assigning a topic number to each comment;
Comment classification means with position information for classifying a plurality of comments with position information for each map route of the route position storage means included in a predetermined range from the position information given to the comment;
For each map path, in commented positional information of multiple, and maps the route topic number extracting means for extracting a predetermined criterion highest topic number based on localization of to the map route as a map route topic number,
A comment classification server comprising map route comment classification means for classifying comments with position information and comments without position information as comments related to the map route based on the map route topic number of each map route. Position information with comments and the positional information without comment collects a plurality of comments are mixed, a comment classification method of the device for extracting the map route associated comments,
The device is
Each map route has a route position storage unit that stores a plurality of positions on the route,
A first step of classifying a plurality of comments into a plurality of topic groups based on the distribution of word frequencies, and assigning a topic number to each comment;
A second step of classifying a plurality of comments with position information for each map route of the route position storage unit included in a predetermined range from the position information given to the comment;
For each map path, in commented positional information of multiple, a third step of extracting a predetermined criterion highest topic number based on localization of to the map route as a map route topic number,
A comment classification method comprising: a fourth step of classifying comments with location information and comments without location information as comments on the map route based on the map route topic number of each map route.

Images