KR102149217B1 - Writing article evaluation method using color information - Google Patents

Abstract

The present invention relates to a method of evaluating a written writing written by an assessee, such as a learner of a language, for a written writing question presented, and in more detail, the similarity of the morphemes included in the sentences written in the written writing written by the assessee. After analyzing and using the analyzed morphemes, evaluation results such as the similarity between the written texts for the writing question, the similarity between each sentence for the written text, and the similarity between the front and back sentences are calculated, and the evaluation results are calculated. It relates to a composition evaluation method using color information provided to the subject by expressing each sentence differently using the three attributes of color: hue, brightness and saturation.

Description

Writing article evaluation method using color information}

The present invention relates to a method for evaluating a written writing written by an assessee, such as a learner who is learning a language, for a written writing question presented, and in more detail, analyzing the similarity of morphemes included in sentences written in the written writing written by the assessee And, using the analyzed morphemes, the evaluation results such as the similarity between the written texts for the writing question, the similarity between each sentence for the written text, and the similarity between the front and back sentences are calculated, and the evaluation results are colored. This field relates to a writing evaluation method using color information that can be expressed in the output unit of the user interface by expressing the sentence differently as the difference between the three attributes of the color, brightness and saturation.

The importance of automatic evaluation systems in language education is increasing day by day. Writing learning is one of the important learning processes in language learning. When evaluating such writing, it can be said that the evaluation of the consistency of the written content is the basis for the evaluation.

In addition, it is difficult for learners to check and understand their writing level with the evaluation results provided by simple text and scores. Consistency evaluation results of written content should be provided in a way that the learner can understand, and based on the results, it should contain information that can help learners establish their own learning plans.

However, the existing automatic evaluation system does not provide sufficient and intuitive information to learners by providing only the result of evaluation by text and score, so continuous research and development is required to solve this problem.

The following is a typical prior art for a system for evaluating written writing.

Republic of Korea Patent Publication No. 10-0848260 relates to a writing learning guidance method and system, the step of constructing a writing learning guidance database including an index database storing index information that is learning information provided to learners, the writing learning guidance database Providing a learning topic to the learner with reference to, receiving writing contents related to the learning topic from the learner, receiving correction information on the writing contents from the instructor after providing the writing contents to the instructor And extracting the index information from the index database included in the writing learning guidance database, and providing teaching information including the correction information and the index information to the learner.

In addition, the prior art was able to obtain the effect of improving the practical writing ability by improving effective communication ability based on the ability to use grammar integrated into writing and one. Since there is an inconvenience of performing corrections on the subject, continuous research and development is required to solve this problem.

Korean Patent Publication No. 10-0848260 (2008.07.18.) Republic of Korea Patent Publication No. 10-2010-0111456 (2010.10.15.) Republic of Korea Patent Publication No. 10-2014-0051606 (2014.05.02.)

The present invention is a technology conceived to improve the problems of the prior art of the writing evaluation system. In the case of the conventional writing evaluation system, in the case of a semi-automatic evaluation system, an evaluator who is a manager manually evaluates the writing, and the evaluation result is computerized. Since it was done in the way of inputting into, there was a problem that it took a long time to evaluate the writing and the cost of labor was high;

In the case of the automatic evaluation system, since the evaluation of human creative writing is performed by the evaluation data randomly inputted into the computer system, it is not possible to accurately evaluate the overall consistency of the writing for the writing question, the main point of the writing evaluation. There was a problem;

In particular, in both the semi-automated evaluation system and the automatic evaluation system, the evaluation result of written writing can be output only as text and score, so that it is difficult for the evaluator and the evaluator, who are administrators, to intuitively check the evaluation result, so that a solution to this The main purpose is to provide.

The present invention is to realize the desired object as described above,

An evaluation request step (S100) in which a user terminal provides a composition written for at least one composition question among a plurality of composition questions included in the question data previously stored in the database by the user terminal to the composition evaluation system;

The writing evaluation system evaluates the sentences included in the writing text provided in the evaluation request step (S100), and expresses the sentences included in the writing text differently by using the three attributes of color: hue, brightness, and saturation according to the evaluation result. An evaluation step (S200) of providing evaluation result sentences to a user terminal; and a composition evaluation method using color information configured including.

Since the composition evaluation method using color information according to the present invention presented as described above can automatically evaluate the composition of the subject for the composition question without direct human involvement, the time required for the evaluation of the composition is short, and it is possible to secure evaluation personnel. It is possible to obtain the effect of reducing labor costs for;

In addition, since it is possible to evaluate the written text by individually evaluating the sentences included in the written text and comparing the entire written text made by the sentences with the writing questions, it is possible to accurately evaluate the consistency of the entire written writing to the writing question. Can achieve the effect;

In particular, since the evaluation result of the written text can be expressed as the difference in color, brightness, and saturation in the written writing itself implemented on the display, the evaluator and the subject can obtain the effect that the evaluation result can be intuitively confirmed through the naked eye. .

1 is a block diagram showing the configuration of a composition evaluation system for implementing a composition evaluation method using color information according to a preferred embodiment of the present invention.
2 is a block diagram showing data included in a database of a device implementing a composition evaluation method using color information according to a preferred embodiment of the present invention.
3 is a block diagram showing an operation procedure of a composition evaluation method using color information according to a preferred embodiment of the present invention.
4 is a block diagram showing an evaluation request step of a composition evaluation method using color information according to a preferred embodiment of the present invention.
5 is a block diagram showing an evaluation step of a composition evaluation method using color information according to a preferred embodiment of the present invention.
6 is a block diagram showing a morpheme grouping step of a composition evaluation method using color information according to a preferred embodiment of the present invention.
7 is a block diagram showing a similarity calculation process of a composition evaluation method using color information according to a preferred embodiment of the present invention.
8 is an exemplary view in which morpheme values are assigned to morphemes in a composition evaluation method using color information according to a preferred embodiment of the present invention.
9 is an exemplary view of grouping morphemes to which morpheme values are assigned in a composition evaluation method using color information according to a preferred embodiment of the present invention.
10 is an exemplary diagram of a sentence value generated for each sentence in a composition evaluation method using color information according to a preferred embodiment of the present invention.
11 is an exemplary view of an evaluation result sentence of a composition evaluation method using color information according to a preferred embodiment of the present invention.

The present invention relates to a method for evaluating written writing for a written question presented by a subject, such as a learner of language learning, wherein the subject responds to any one or more of a plurality of writing questions included in question data pre-stored in a database. An evaluation request step (S100) in which the user terminal provides the written writing to the writing evaluation system;

The writing evaluation system evaluates the sentences included in the writing text provided in the evaluation request step (S100), and expresses the sentences constituting the writing text differently using the three attributes of color: hue, brightness, and saturation according to the evaluation result. It relates to a composition evaluation method using color information including; evaluation step (S200) of providing evaluation result sentences to a user terminal.

Hereinafter, the present invention will be described in detail with reference to FIGS. 1 to 11 showing an embodiment of the present invention.

First, the composition evaluation method of the present invention is characterized in that the composition evaluation system and the user terminal are implemented on a hardware basis.

The writing evaluation system may be a known computer as a server device, and the user terminal may be a well-known smartphone, and the user terminal performs a communication network with the writing evaluation system, and allows the subject to check writing questions. , It is characterized in that a dedicated application that performs a function of transmitting the written writing to the writing system is mounted.

In addition, the composition evaluation system is characterized in that it is configured to include at least an input unit, an evaluation unit, a control unit, and a database as shown in FIG.

In addition, the composition evaluation system according to the present invention is macroscopically configured to include an evaluation request step (S100) and an evaluation step (S200) as shown in FIG.

The evaluation request step (S100) is a step in which the user terminal provides the writing written for any one or more writing questions among a plurality of writing questions included in the question data pre-stored in the database to the writing evaluation system. This is the process of providing the target writing text to the writing evaluation system through the user terminal.

At this time, the writing is written by the evaluator based on the question data previously stored in the database of the writing evaluation system, and the question data is data previously stored in the database by the manager (evaluator) before evaluation, and a plurality of writing questions It has a configuration that includes.

Specifically, when a manager inputs a plurality of writing questions through the input unit of the writing evaluation system, the inputted writing questions are stored in a database as question data, and at the time of evaluation, any one or more of the plurality of writing questions stored in the database The question is provided to the user terminal so that the subject can check the writing question.

Specifically, in the evaluation request step (S100), a writing questionnaire providing step (S110) in which the writing evaluation system provides a writing questionnaire among a plurality of writing questions included in the question data to a user terminal and the provided writing question And a writing writing providing step (S120) in which the user terminal provides the written writing written by the assessee to the writing evaluation system, and a writing writing storing step (S130) in which the provided writing writing is stored in the database by the writing evaluation system (S130). .

That is, the evaluation request step (S100) has a configuration including a macroscopically a writing question providing step (S110), a writing writing providing step (S120), and a writing writing storing step (S130).

Specifically, in the writing question provision step (S110), the control unit of the writing evaluation system extracts any one or more writing questions from among a plurality of writing questions previously stored as question data in a database, and provides the extracted writing questions to the user terminal. As a process, the extraction of the writing question may be an extraction selected by an administrator or an extraction arbitrarily selected by the control unit.

The writing question extracted in the writing question providing step (S110) and provided to the user terminal of the assessee may be a single writing question or a plurality of writing questions according to the purpose of the writing evaluation.

The writing writing providing step (S120) is a process in which the evaluator checks the writing question and provides the writing written according to the checked writing question to the writing evaluation system by the user terminal, and the evaluator checks the writing question through the user terminal. And write the writing according to the confirmed writing questions.

The user terminal is equipped with a dedicated application, and the application provides a user interface that allows the subject (terminal user) to check the writing questions provided by the writing evaluation system, and allows the user to write a writing according to the checked writing questions. It provides a user interface and transmits the written writing to the writing evaluation system through a wired or wireless communication network.

The writing writing storage step S130 is a process in which the control unit stores the writing text provided from the user terminal through the evaluation request step S120 as writing writing data in the database.

In the evaluation step (S200), the evaluation unit of the writing evaluation system evaluates the sentences constituting the writing text provided through the evaluation request step (S100), and according to the evaluation result, the sentences constituting the writing text are selected as three attributes of color. , Providing the evaluation result sentences differently expressed using brightness and saturation to the user terminal, which evaluates the similarity between the sentence constituting the composition and the words (morphemes) included in the sentence, and color and brightness according to the evaluation result. This is the step of generating an evaluation result sentence that expresses differently the sentences constituting the composition text to be evaluated using saturation.

In other words, the evaluation step (S200) compares the sentences and words included in the written writing with the writing questions, and the similarity of the whole writing to the writing question, the similarity between each sentence of the written writing, and between the front and back sentences. It is characterized in that it evaluates the degree of similarity of and, according to the evaluation result, generates evaluation result sentences in which three attributes of color, such as hue, brightness, and saturation, are differently displayed and provided to the user terminal.

Preferably, in the evaluation step (S200), a plurality of morphemes included in the composition question and the written text are separated, the separated morphemes are grouped among similar morphemes, and morphemes included in the same morpheme group are treated as the same morphemes. A morpheme grouping step (S210) of giving the same morpheme value, which is a standard morpheme value; Sentence value calculation step of generating sentence values for each sentence by applying the standard morpheme values to the morphemes described in each of the writing questions and the plurality of sentences included in the written text, and summing the standard morpheme values included in each sentence ( S220); Using each of the sentence values of the writing question and the plurality of sentences included in the written article, i) the first degree of similarity, which is the degree of similarity between the written articles for the composition question, ii) the second degree of similarity, which is the degree of similarity between each sentence for the written article, and Iii) a similarity calculation step (S230) of calculating a third similarity degree, which is a degree of similarity between the front sentence and the back sentence; Color values for generating color values, brightness values, and saturation values representing the color, brightness, and saturation of each sentence for reflecting the first, second, and third similarities to each of a plurality of sentences included in a written article Calculating step (S240) and; An evaluation result sentence generation step (S250) of generating an evaluation result sentence in which the generated color value, brightness value, and saturation value are reflected in each sentence; It characterized in that it comprises a; evaluation result transmission step (S260) of providing the generated evaluation result sentence to the user terminal.

That is, the evaluation step (S200) of the above configuration macroscopically morpheme grouping step (S210), sentence value calculation step (S220), similarity calculation step (S230), color value calculation step (S240), evaluation result sentence generation step (S250) And a step of transmitting the evaluation result (S260).

Specifically, in the morpheme grouping step (S210), a plurality of morphemes included in the composition question and written text are separated, the separated morphemes are grouped among similar morphemes, and morphemes included in the same morpheme group are treated as the same morphemes. This is the process of assigning the standard morpheme value, which is the same morpheme value, by analyzing the morphemes included in the composition question and the written text.

In this case, in the morpheme grouping step (S210), a plurality of morphemes included in a composition question and a written article are separated, and a plurality of morpheme groups are formed by grouping the separated morphemes with similar meanings, and the plurality of morphemes Among the groups, morphemes included in the same morpheme group can be treated as the same morpheme to give the same morpheme value, which is a standard morpheme value, and can be processed in various ways.

Preferably, the database stores morpheme data to which morpheme values expressed as coordinates are assigned to each of a plurality of morphemes, and the morpheme grouping step (S210) extracts the morpheme data and writing text data stored in the database, and the A morpheme separation processing step (S211) of separating morphemes included in the writing by using morpheme data and giving the morpheme value to each of the separated morphemes; A morpheme grouping processing step (S212) of grouping a plurality of morphemes to which a morpheme value is assigned in the morpheme separation processing step (S211) with morphemes having coordinates of a similar region; And a standard morpheme value assignment processing step (S213) of assigning a standard morpheme value that is the same morpheme value to morphemes included in the same morpheme group among the morphemes grouped in the morpheme grouping processing step (S212).

That is, the morpheme grouping step (S210) of the above configuration has a configuration macroscopically including a morpheme separation processing step (S211), a morpheme grouping processing step (S212), and a standard morpheme value assignment processing step (S213).

Specifically, in the morpheme separation processing step (S211), the evaluation unit extracts the morpheme data and the writing text data from the database, separates the morphemes included in the writing text using the morpheme data, and separates the morphemes into each of the separated morphemes. As a process of assigning the morpheme value, a process of separating the morphemes described in the writing question and the written writing based on the morpheme list included in the morpheme data previously stored in the database by the administrator.

That is, the morpheme data may include a morpheme list, a morpheme group list, a morpheme value list, and a standard morpheme value list.

In this case, the morpheme list is a configuration that acts like a morpheme dictionary, and includes a plurality of general morphemes, and provides basic data for extracting and separating morphemes included in a composition question and a written article, and the morpheme group list Is basic data obtained by grouping morphemes included in the morpheme list with those having similar meanings, the morpheme value list is basic data in which specific coordinates are assigned to each of the morphemes included in the morpheme list, and the standard morpheme value list Is basic data including single morpheme values assigned to each of a plurality of morpheme groups included in the morpheme group list.

In connection with the above, the morpheme separation processing step (S211) separates the morphemes included in the composition question and the writing text using the morpheme list included in the morpheme data, and matches each of the separated morphemes to the morpheme value list This is a process of assigning a morpheme value expressed as a coordinate to each separated morpheme.

In addition, the morpheme grouping processing step (S212) is a process of grouping a plurality of morphemes to which a morpheme value is assigned in the morpheme separation processing step (S211) with morphemes having coordinates of a similar region, and the evaluation unit uses the morpheme group list. Thus, morphemes having the same or similar meaning are grouped to belong to the same morpheme group.

In addition, the standard morpheme value assignment processing step (S213) is a process of assigning a standard morpheme value that is the same morpheme value to morphemes included in the same morpheme group among the morphemes grouped in the morpheme grouping processing step (S212). This is a process of assigning the same standard morpheme value to each of the morphemes having the same or similar meaning belonging to the same morpheme group by using the standard morpheme value list.

That is, the composition question analyzed in the morpheme grouping step S210 is extracted as morphemes included in the composition question in sentence form, and the extracted morphemes are grouped with those having the same or similar meaning. At this time, the database is configured to include morpheme data (morpheme group list data) in which morphemes having the same or similar meaning are grouped, and the morpheme data previously stored in the database is retrieved to the evaluation unit in the morpheme grouping step (S210). The extracted morphemes of the composition question can be matched to a plurality of morpheme groups, and group information of the morphemes of the composition question matched to each morpheme group is stored in a database.

In addition, if the morpheme group list data of the morpheme data belongs to the same group of a plurality of morphemes having the same or similar meaning, and morphemes having different meanings belong to different groups, various data The morpheme data may be stored in a database in a processing method, and preferably, the morpheme data may be configured in a data processing method in which a morpheme value expressed as a coordinate is assigned to each of a plurality of morphemes and previously stored in the database.

That is, the morpheme value list data of the morpheme data of the configuration has a configuration in which morpheme values having different coordinates (coordinates expressed as (x, y) in one embodiment) are assigned according to the meaning of each of the plurality of morphemes. Specifically, morphemes with the same meaning have the same coordinates and are given the same morpheme value, and morphemes with similar meanings have coordinates of a position where one morpheme and another similar morpheme are adjacent, so that the similar morpheme values are each The morphemes of different meanings have the coordinates of a distant position between one morpheme and the other morpheme, so that different morpheme values can be respectively assigned.

In addition, morphemes are also extracted in the same manner as the composition question in the composition text analyzed in the morpheme grouping step (S210). That is, the written text is extracted as morphemes included in the written writing in the form of sentences, and the extracted morphemes are grouped with those having the same or similar meaning. At this time, the morphemes included in each of the plurality of sentences included in the written text are configured to be able to classify the morphemes included in the corresponding sentence.

As described above, after the morphemes included in the composition question and the written text are grouped into a plurality of morpheme groups, the morphemes included in the same morpheme group are treated as the same morphemes, and a standard morpheme value, which is the same morpheme value, is assigned.

That is, the standard morpheme value is included in the standard morpheme value list data, and is a single morpheme value assigned to each of a plurality of morpheme groups, and each of the morphemes having the same or similar meaning included in the same morpheme group is as described above. Although they have unique morpheme values that have the same or similar coordinates, in writing evaluation, since multiple morphemes included in the same morpheme group are interpreted as the same meaning and must have the same evaluation result, the standard morpheme value is the same morpheme group. The same evaluation result is used as a factor for reflecting the same evaluation result on a plurality of morphemes included in

In the sentence value calculation step (S220), the standard morpheme values are respectively applied to the morphemes described in each of the plurality of sentences included in the composition question and the written article, and the standard morpheme values included in each sentence of the composition question and the written article are As a process of generating sentence values of each sentence constituting a composition question and a written article by summing up, the plurality of morpheme groups are separated into a plurality of morpheme groups in the morpheme grouping step (S210), and a standard morpheme value of any one morpheme group is assigned. This is a process of generating sentence values for each of the plurality of sentences constituting the composition question and the written article after matching the morphemes of in the composition question and the sentences included in the composition article.

That is, in the sentence value calculation step (S220), the standard morpheme values of each of the morphemes included in a single or a plurality of sentences included in each of the writing question and the written text are summed for each sentence, This is the process of generating sentence values that are the sum of morpheme values for each sentence.

In more detail, a plurality of morphemes are included in any one sentence included in the written text, and in this case, each of the plurality of morphemes included in the one sentence (hereinafter referred to as'one embodiment sentence') is each other. Each morpheme belonging to the same or different morpheme groups may belong to the same or different morpheme groups, and each morpheme belonging to the same or different morpheme groups has the same or different standard morpheme values, and the standard morpheme values of the plurality of morphemes included in the sentence of one embodiment are summed to perform a corresponding one It is created as the sentence value of the sentence.

In connection with the above, when the morpheme value is expressed as coordinates as described above, the standard morpheme value may also be expressed as coordinates, and the sentence value may also be expressed as coordinates. That is, the sentence value may be expressed as coordinates by adding coordinate values of the same axis of the standard morpheme value expressed as coordinates.

The similarity calculation step (S230) uses each of the sentence values of the composition question and the sentence values of each sentence constituting the composition, i) a first degree of similarity, which is the degree of similarity between the composition question and the composition article, ii) constructs a composition article. It is a process of calculating the second similarity, which is the degree of similarity between each sentence, and iii) the third similarity, which is the degree of similarity between one adjacent preceding sentence and the following sentence among a plurality of sentences constituting the composition. This is the process of evaluating the completeness of writing.

That is, in the similarity calculation step (S230), using the sentence value generated in the sentence value calculation step (S220), the consistency of the written writing to the writing question can be evaluated as the first similarity, and Consistency of each sentence with respect to the entire sentence may be evaluated as the second degree of similarity, and consistency between sentences adjacent to each other among a plurality of sentences constituting a written article may be evaluated as the third degree of similarity.

At this time, the first similarity, the second similarity, and the third similarity are evaluated as having higher consistency as they approach -1 or 1, and as opposed to the above, the closer they are to 0, the lower their consistency.

Specifically, in the similarity calculation step (S230), the first degree of similarity, the second degree of similarity, and the third degree of similarity are calculated using the sentence values.

Preferably, the similarity calculation step (S230) includes a cosine derived value generation processing step (S231) of generating a cosine derived value of the writing question by applying a cosine function to the sentence value of the writing question; A cosine derivation value generation processing step (S232) for generating a cosine derivation value of each of the plurality of sentences constituting the composition by applying a cosine function to the sentence values of each of the plurality of sentences constituting the written article; And a similarity calculation processing step (S233) of calculating the first, second, and third similarities, respectively, using the cosine derivation value of the writing question and the cosine derivation value of sentences constituting the written writing. can do.

That is, the similarity calculation step (S230) has a configuration including macroscopically a cosine derived value generation processing step (S231), a cosine derived value generation processing step of sentences (S232), and a similarity calculation processing step (S233). .

At this time, the cosine derivation value generation processing step (S231) of the composition question is a process of generating a cosine derivation value of the composition question having a value in the range of -1 to +1 by applying a cosine function to the sentence value of the composition question. , This is a process of generating a standard value for evaluation by using the sentence value of the writing question, which is the evaluation criterion for written writing.

In connection with the above, the cosine derivation value generation processing step (S231) of the composition question applies a cosine function to the sentence value of the composition question expressed in the form of coordinates, so that the cosine of the composition question has a value in the range of -1 to +1. A derivation value is generated. A composition article containing morphemes having the same or similar meaning as the meaning of the morphemes included in the composition question and the sentences constituting the writing article are given a high evaluation score (expressed in color information.) Compositions that contain morphemes with meanings that are different from the meanings of the morphemes included in the question and the sentences that compose the writings are structured so that low evaluation scores (expressed in color information) can be given.

At this time, since the cosine derivation value of the composition question is calculated by applying the cosine function to the sentence value of the composition question expressed in the form of coordinates, it is only a value within the range of -1 to +1 unless a constant is given to the cosine function. Is calculated. In addition, since the present invention designates the cosine derivation value of the composition question as the reference value, the absolute value 1, the closer the cosine derivation value of each sentence constituting the composition of the composition is to the absolute value 1, the higher the similarity to the composition question is, and The closer it is, the lower the similarity to the writing question can be interpreted.

In addition, as the cosine function, a cosine function to which an appropriate function is applied may be used according to judgment of a person skilled in the art. Hereinafter, a case where the x-coordinate and y-coordinate of the sentence value of the composition question expressed in the form of coordinates used as a variable of the cosine function are expressed as coordinates based on 0, which is the origin of the x-y axis, will be described as a preferred embodiment.

In addition, the cosine derivation value generation processing step (S232) of the sentences is a process of generating each cosine derivation value of the sentences included in the composition by applying the cosine function to each sentence value of a plurality of sentences constituting the written article. , This is a process of calculating sentence values of sentences expressed in coordinates included in the writing text as a specific single evaluation value.

That is, in the cosine derivation value generation processing step (S232) of the sentences, the cosine derivation value generation processing step (S231) of the writing question is used to specify the sentence values of each of the sentences constituting the writing text. By calculating the evaluation value, it is configured to be able to compare the cosine derivation value of the writing question and the cosine derivation value of the sentences constituting the written writing.

In addition, the similarity calculation processing step (S233) is a process of calculating the first degree of similarity, the second degree of similarity, and the third degree of similarity, respectively, using the cosine derivation value of the composition question and the cosine derivation value of each sentence constituting the written article. This is a process of evaluating the similarity of the writing to the writing question, the similarity of each sentence constituting the writing to the entire composition, and the similarity between the preceding sentence and the latter sentence adjacent to the sentences constituting the composition. .

That is, the first degree of similarity is a value representing the degree of similarity between a composition question and a written article, and the composition average cosine derivation value calculated by summing and averaging the cosine derivation values of each sentence included in the composition article and the composition It can be generated by comparing the cosine derived values of the questions.

At this time, the composition average cosine derivation value represents the average of the cosine derivation values of all sentences included in the composition article, and the cosine derivation value of the summed sentences after summing the cosine derivation values of all sentences included in the composition article Dividing the total of by the number of sentences; It is also possible to calculate the average cosine derivation value of the written writing by summing the sentence values of all sentences included in the written text and applying the cosine function to the total sentence values of the written written text calculated by dividing the summed sentence values by the number of sentences.

Specifically, by comparing the average cosine derivation value of the written writing and the cosine derivation value of the writing question, the closer the absolute value of the average cosine derivation value of the written writing is to the absolute value of the cosine derivation value of the writing question, the more Contrary to the above, the closer the absolute value of the average cosine derivation value of the written writing is to 0, it is evaluated that it indicates that the written writing is not properly written in the writing question.

In addition, the second degree of similarity is a value representing the degree of similarity between the written text and each sentence constituting the written text, and is calculated by summing and averaging the cosine induction values of each of the sentences included in the written text. It may be generated by comparing the value and the cosine derivation value of each of a plurality of sentences constituting the written text.

In this case, the average cosine derivation value of the written writing may be calculated in the same manner as the average cosine derivation value of the written writing calculated when the first similarity is calculated, and the cosine derivation value of each sentence is a plurality of It represents the cosine derived value of each sentence.

That is, the second degree of similarity compares the cosine derivation value of each sentence constituting the written article with the average cosine derivation value of the composition article, and evaluates how consistently each sentence constituting the composition article is written within the composition article. Configuration.

In addition, the second similarity is generated by a number corresponding to the number of a plurality of sentences constituting a written article.

In addition, the third similarity is a value representing the degree of similarity between a preceding sentence and a subsequent sentence among a plurality of sentences included in the composition, and is a value confirming the organic connection between the preceding sentence and the latter sentence.

That is, the third similarity is the cosine derivation value of the first sentence corresponding to any one preceding sentence among the cosine derivation values of each sentence of a plurality of sentences included in the composition and the subsequent sentence of the preceding sentence. By comparing the cosine derivation values of the corresponding second sentence with each other, the organic connection between adjacent sentences can be confirmed.

In addition, the color value calculation step (S240) includes a color value representing the color, brightness, and saturation of each sentence for reflecting the first, second, and third similarities to each of a plurality of sentences included in the composition, As a process of generating a brightness value and a saturation value, color information is reflected in a written article using the first degree of similarity, the second degree of similarity, and the third degree of similarity generated in the similarity calculation step S230.

That is, each of the first, second, and third similarities calculated through the similarity calculation step (S230) is i) the first similarity, which is the degree of similarity between the writing question and the written text, ii) each constituting the writing A second degree of similarity, which is the degree of similarity between sentences, and iii) a third degree of similarity, which is the degree of similarity between one of the plurality of sentences that are adjacent to each other, among the plurality of sentences constituting the written text. Each is reflected as color information.

In this case, the color information is a color value, a brightness value, and a saturation value representing hue, brightness, and saturation, and the color information of each sentence constituting the writing is designated as the color value, brightness value, and saturation value, and the following evaluation In the result sentence generation step (S250), an evaluation result sentence, which is an evaluation result of a written article composed of a plurality of sentences expressed in different color information, is generated. Of course, sentences having the same degree of similarity (first degree of likeness, second degree of similarity, and third degree of similarity) among a plurality of sentences included in the written article have the same color information.

In connection with the above, in the color value calculation step (S240), specific color information may be designated according to the similarities of each of a plurality of sentences included in the composition, and may be processed in various ways, preferably known. The color information can be processed to be designated by the HSV (Hue, Saturation, Value) color space model.

That is, the HSV (Hue, Saturation, Value) color space model is expressed as a three-dimensional figure in the shape of a cylinder or an inverted cone, and the hue, saturation, and saturation (Value) are specified within the three-dimensional figure. It is a model specified by coordinates.

That is, a specific color value, a brightness value, and a saturation value representing a specific hue, saturation, and saturation value are designated as specific coordinate values on the HSV color space model, and the data related to this is the composition of the present invention. It has a state previously stored as standard color information data in the database of the evaluation system.

In this case, the color value calculation step (S240) is a specific color value and brightness value that can be implemented in the HSV color space model corresponding to the first, second, and third similarities of each of the plurality of sentences constituting the written writing. And a process of generating a saturation value, wherein the hue value, brightness value, and saturation value are values for expressing an evaluation result of a written article as color information.

Specifically, the first degree of similarity is a value calculated by comparing the average cosine derivation value of the composition article calculated by summing and averaging the cosine derivation values of each of the plurality of sentences constituting the composition article and the cosine derivation value of the composition question. As a result, the brightness value in the HSV (Hue, Saturation, Value) color space model is determined according to the degree of similarity between the cosine derivation value of the writing question and the average cosine derivation value of the written writing, and the calculated first similarity and corresponding Corresponding relationship information of the brightness value has a state previously stored in the database of the composition evaluation system of the present invention.

For example, if the cosine derivation value of the composition question is 0.9 and the average cosine derivation value of the written article is 0.85, the first degree of similarity is 0.85, which is the average cosine derivation value of the composition when the cosine derivation value of the composition question is 1 The relative value of is the first degree of similarity.

That is, (0.9:0.85=1:0.944) is established by the proportional expression, so that the first similarity is calculated as 0.944, and the brightness value corresponding to 0.944 is the correspondence of the first similarity and the corresponding brightness value stored in the database. It is determined by relationship information, and the brightness value corresponding to the first similarity degree can be arbitrarily set by the user and stored in the database.

As a result, the brightness of the written text is expressed on the evaluation result sentence with the determined brightness value.

In addition, the second degree of similarity is calculated by summing and averaging the cosine derivation values of each of the plurality of sentences constituting the written article, and comparing the cosine derivation value of each of the plurality of sentences constituting the written article. As a calculated value, according to the degree of similarity between the comparison result, the average cosine derivation value of the written writing and the cosine derivation value of each of the plurality of sentences constituting the written writing, the color in the HSV (Hue, Saturation, Value) color space model The value is determined, and the calculated second similarity and the corresponding relationship information of the color value corresponding thereto have a state previously stored in the database of the composition evaluation system of the present invention.

For example, when the average cosine derivation value of a written article is 0.9 and the cosine derivation value of any one sentence among a plurality of sentences constituting the written article is 0.7, the second similarity is when the average cosine derivation value of the written article is 1 A relative value of 0.7, which is a cosine derived value of any one of the plurality of sentences constituting the written writing of, becomes the second similarity.

That is, (0.9:0.7=1:0.78) relationship is established by the proportional expression, and the second similarity is calculated as 0.78, and the color value corresponding to 0.78 is the correspondence of the second similarity stored in the database and the corresponding color value. The color value determined by relationship information and corresponding to the second similarity degree may be arbitrarily set by the user and stored in the database.

As a result, the color of any one sentence among the plurality of sentences constituting the written text is expressed on the evaluation result sentence with the determined color value.

In addition, the third similarity is the cosine derivation value of the first sentence corresponding to any one of the cosine derivation values of each of the plurality of sentences constituting the written writing and the second sentence corresponding to the rear sentence of the preceding sentence. A value calculated by comparing the cosine derivation value of, as a result of the comparison, the degree of similarity between the cosine derivation value of the first sentence corresponding to any one preceding sentence and the cosine derivation value of the second sentence corresponding to the rear sentence of the preceding sentence. According to this, the saturation value in the HSV (Hue, Saturation, Value) color space model is determined, and the calculated third similarity and correspondence relationship information of the corresponding saturation value are previously stored in the database of the writing evaluation system of the present invention. Have.

For example, if the cosine derivation value of the first sentence corresponding to the previous sentence is 0.9 and the cosine derivation value of the second sentence corresponding to the latter sentence is 0.5, the third similarity is the cosine of the first sentence corresponding to the previous sentence. When the derived value of 0.9 is standardized as 1, the relative value of 0.5, which is the cosine derived value of the second sentence corresponding to the latter sentence, becomes the third similarity.

That is, (0.9:0.5=1:0.56) relationship is established by the proportional expression, the third similarity is calculated as 0.56, and the saturation value corresponding to 0.56 is the correspondence of the third similarity stored in the database and the corresponding saturation value. The saturation value, which is determined by the relationship information, and corresponding to the third similarity degree, can be arbitrarily set by the user and stored in the database.

As a result, the saturation of the latter sentence of any one of the plurality of sentences constituting the written text is expressed on the evaluation result sentence by the determined saturation value.

In addition, in the color value calculation step (S240), a color value, a brightness value, and a saturation value for the writing question may be generated, and the evaluation unit randomly designates the color value, brightness value, and saturation value for the composition question, or You can also specify.

In addition, color value data representing color values, brightness values, and saturation values of each of the writing questions generated in the color value calculation step (S240) and a plurality of sentences constituting the writing text, brightness value data, and saturation value data are color information data. As stored in the database.

In addition, the evaluation result sentence generation step (S250) is a process of generating an evaluation result sentence in which the color value, brightness value, and saturation value are reflected in each of the evaluation subject sentences, and the evaluation unit is the evaluation subject composition text and color value stored in the database. , Color value data representing brightness values and saturation values, color information data including brightness value data and saturation value data are extracted, and corresponding color values, brightness values, and texts are applied to each sentence constituting the extracted writing. The evaluation result sentences that are displayed by reflecting the saturation value are generated.

In addition, the evaluation result transmission step (S260) is a process in which the generated evaluation result sentences are provided to the user terminal as evaluation results, and the evaluator checks the evaluation result sentences provided by displaying each sentence in different color information, You will be able to check the evaluation results of a written writing.

The following describes an example in which an evaluator receives a composition evaluation by using the composition evaluation method using color information according to the present invention.

First, the control unit of the writing evaluation system provides a writing question “Describe a memorable trip” among a plurality of writing questions previously stored in the database to the user terminal.

Thereafter, the assessee who checks the writing question writes the following writing using the input interface of the user terminal.

'I went to Korea in May of this year, and it was my first trip abroad.

I went to Korea alone because I wanted to see a drama and go to a place where I film a drama.

But I didn't know where the drama location cafe street was.

So, I called a friend who knows where it is, and I visited the Jukjeon Cafe Street with a friend who was learning Korean and living in Korea.

It was too far from Sinchon to Jukjeon Station, so I changed subways.

It was like taking the subway for 2 hours and walking about 20 minutes.

This is really beautiful.

There are many people, but the streets are clean, so I wish it snowed.’

The composition written by the subject and provided to the composition evaluation system by the user terminal is stored in a database by the control unit of the composition evaluation system as composition writing data.

Thereafter, the evaluation unit of the composition evaluation system processes the morpheme grouping step (S210). That is, the evaluation unit separates the morphemes included in the composition as follows, using the morpheme list data previously stored in the database.

‘Memory/E/M/E/Travel/E/Daeha/Female/ Narrative/Ha/Si/O/

This year/ May/ To/ Korea/ To/ Go/ Ah/ Bo/ I/ I/ I/ First/ Foreign/ Travel/ I/ I/ I/.

Drama/ Watching/ Watching/ Drama/ Filming/ Place/ To/ Watch/ Want/ Because/ Alone/ Korea/ To/ I went/.

But/ Drama/ Location/ Cafe/ Street/ A/ Where/ This/ B know/ Don't know/ Don't know/.

So/ Where/이/ㄴ지/ Know/A/ Friend/ To/ Phone/ Ha/ Did/ But/ Korea/ From/ Korean/ Actor/ High/ Sal/ A/

Friends/ with/ with/ at Jukjeon cafe street/ at/ watching/ha/ha/ha/.

From Sinchon/To/To Jukjeon Station/Too/Too/Moon/Come on/Subway/E/Go/A/Take/Go/Go/A/T/D/Go/.

2/ Hours/ During/ Subway/ E/ T/ W/ Would/ Same/ High/ 20/ Minute/ About/ Walk/ W/ W/ W/ W/ W/.

Here/Go/ Real/ Beautiful/ Beautiful/.

People/s/do/many/but/ street/go/ clean/ha/nade/ snow/i/oh/if/ good/will/do/'

Thereafter, the evaluation unit assigns morpheme values expressed in coordinates of (x, y) to each of the morphemes using the morpheme value list data previously stored in the database (see Fig. 8), and stores the morpheme group list data previously stored in the database. The morphemes to which the morpheme values are assigned are grouped by using (see FIG. 9).

Thereafter, the evaluation unit assigns a standard morpheme value of the corresponding morpheme group to morphemes belonging to each of the plurality of morpheme groups by using the standard morpheme value list data previously stored in the database.

Thereafter, the evaluation unit processes the sentence value calculation step (S220), applies the standard morpheme values to each of the morphemes described in each of the writing questions and the plurality of sentences included in the written article, and applies the standard morpheme values included in each sentence. The summation is performed to generate the sentence value of each sentence (see FIG. 10). At this time, the evaluation unit also calculates the average of the sentence values of the written text by summing the sentence values of all the sentences included in the written text.

Thereafter, the evaluation unit processes the similarity calculation step (S230), calculates a first degree of similarity, a second degree of similarity, and a third degree of similarity, and stores the calculated first degree of similarity, the second degree of similarity, and the third degree of similarity in the database.

Thereafter, the evaluation unit performs a color value calculation step (S240) of generating a color value, a brightness value, and a saturation value to be reflected in each sentence.

That is, the evaluation unit generates a color value, a lightness value, and a saturation value representing the difference in brightness according to the first similarity, the difference in color according to the second similarity, and the difference in saturation according to the third similarity, for each sentence.

Thereafter, the evaluation unit stores color value data representing color values, brightness values, and saturation values of each of the plurality of sentences generated in the color value calculation step S240, brightness value data, and saturation value data as color information data in a database, and , An evaluation result sentence is generated by reflecting the generated color value, brightness value, and saturation value for each sentence in each sentence.

Thereafter, the generated evaluation result sentence is transmitted to the user terminal.

As shown in FIG. 11, an evaluator, who is a user of the user terminal, can check the evaluation result sentence displayed with color information.

The above has been described with reference to a preferred embodiment of the present invention, and is not limited to the above embodiment, and through the above embodiment, those skilled in the art to which the present invention belongs do not depart from the gist of the present invention. This can be done with various changes in.

S100: Evaluation request step
S110: The stage of providing writing questions
S120: Evaluation request step
S130: Writing writing saving step
S200: Evaluation stage
S210: morpheme grouping step
S211: Morphological separation processing step
S212: morpheme grouping processing step
S213: Standard morpheme value assignment processing step
S220: sentence value calculation step
S230: Similarity calculation step
S231: Cosine derived value generation processing step of the writing question
S232: Processing step of generating cosine derived values of sentences
S233: Similarity calculation processing step
S240: Color value calculation step
S250: Evaluation result sentence generation step
S260: Evaluation result transmission step

Claims (8)

In the writing evaluation method using color information,
An evaluation request step (S100) in which a user terminal provides a composition written for at least one composition question among a plurality of composition questions included in the question data previously stored in the database of the composition evaluation system by the user terminal to the composition evaluation system;
The writing evaluation system evaluates the sentences included in the writing text provided in the evaluation request step (S100), and expresses the sentences constituting the writing text differently using the three attributes of color: hue, brightness, and saturation according to the evaluation result. Including; an evaluation step (S200) of providing the evaluation result sentences to the user terminal,

The evaluation step (S200),
The composition question and a plurality of morphemes included in the written text are separated, the separated morphemes are grouped among similar morphemes, and the morphemes included in the same morpheme group are treated as the same morphemes to give the same morpheme value, a standard morpheme value. Morpheme grouping step (S210);
The standard morpheme values are applied to the morphemes in each of the writing questions and the plurality of sentences included in the written text, and the standard morpheme values included in each sentence of the writing question and the written text are summed to form a writing question and a written text. A sentence value calculation step (S220) of generating a sentence value of each sentence to be executed;
Using each of the sentence values of the writing question and each sentence constituting the written article, i) the first degree of similarity, which is the degree of similarity between the composition question and the written article, ii) the second degree of similarity, which is the degree of similarity between each sentence constituting the written article. And iii) a similarity calculation step (S230) of calculating a third degree of similarity, which is a degree of similarity between a preceding sentence and a subsequent sentence adjacent to each other among a plurality of sentences constituting the written writing.
A color value calculation step (S240) of generating a color value, a brightness value, and a saturation value for reflecting the calculated first, second, and third similarities to each of a plurality of sentences constituting a written article;
An evaluation result sentence generation step (S250) of generating an evaluation result sentence in which the generated color value, brightness value, and saturation value are reflected in each sentence;
Including; an evaluation result transmission step (S260) of providing the generated evaluation result sentence to the user terminal,
The color value, brightness value, and saturation value generated in the color value calculation step (S240) are values for expressing the evaluation result of the composition as color information.
The method of claim 1,
The evaluation request step (S100),
A writing question providing step (S110) in which a writing evaluation system provides a writing evaluation system to a user terminal with at least one writing question among a plurality of writing questions included in the question data,
A writing writing providing step (S120) in which the user terminal provides the writing writing written by the assessee for the provided writing question to the writing evaluation system,
Writing evaluation method using color information, characterized in that it comprises a; writing writing storage step (S130) of storing the provided writing writing in a database by the writing evaluation system.
delete The method of claim 1,
In the database,
It is configured to further include morpheme data in which morpheme values expressed as coordinates are assigned to each of the plurality of morphemes,
The morpheme grouping step (S210),
The morpheme separation processing step of receiving the morpheme data and the writing text data in which the writing text is data from the database, separating the morphemes included in the writing text by using the morpheme data, and giving the morpheme value to each of the separated morphemes (S211);
A morpheme grouping processing step (S212) of grouping a plurality of morphemes to which morpheme values are assigned in the morpheme separation processing step (S211) with morphemes having coordinates of a similar region;
And a standard morpheme value assignment processing step (S213) of assigning a standard morpheme value that is the same morpheme value to morphemes included in the same morpheme group among the morphemes grouped in the morpheme grouping processing step (S212). A method of evaluating writing using color information.
The method of claim 4,
The similarity calculation step (S230),
A cosine derivation value generation processing step (S231) of generating a cosine derivation value of the composition question by applying a cosine function to the sentence value of the composition question (S231);
A cosine derived value generation processing step (S232) of generating a cosine derived value of each of the plurality of sentences constituting the composition by applying a cosine function to each sentence value of the plurality of sentences constituting the written article;
Including a similarity calculation processing step (S233) of calculating the first, second, and third similarities, respectively, using the cosine derivation value of the composition question and the cosine derivation value of each sentence constituting the written article. Composition evaluation method using color information, characterized in that configured.
The method of claim 5,
The first degree of similarity,
It is a value calculated by comparing the average cosine derivation value of the composition article calculated by summing and averaging the cosine derivation values of each of the plurality of sentences constituting the composition article and the cosine derivation value of the writing question
The brightness value in the HSV (Hue, Saturation, Value) color space model is determined according to the calculated first similarity,
The composition evaluation method using color information, characterized in that the relationship information between the first similarity degree and the brightness value on the HSV (Hue, Saturation, Value) color space model is stored in a database of a composition evaluation system.
The method of claim 5,
The second similarity is,
It is a value calculated by comparing the average cosine derivation value of a composition article calculated by summing and averaging the cosine derivation values of each of the plurality of sentences constituting the composition article with the cosine derivation value of each of the plurality of sentences constituting the composition article,
The color value in the HSV (Hue, Saturation, Value) color space model is determined according to the calculated second similarity,
The composition evaluation method using color information, characterized in that the relationship information between the second similarity and the color value in the HSV (Hue, Saturation, Value) color space model is stored in a database of a composition evaluation system.
The method of claim 5,
The third similarity is,
Of the cosine derivation values of each of the plurality of sentences constituting the composition, the cosine derivation value of the first sentence corresponding to any one preceding sentence and the cosine derivation value of the second sentence corresponding to the rear sentence of the previous sentence are compared Is the calculated value,
The saturation value in the HSV (Hue, Saturation, Value) color space model is determined according to the calculated third similarity,
A composition evaluation method using color information, characterized in that the relationship information between the third similarity degree and the saturation value in the HSV (Hue, Saturation, Value) color space model is stored in a database of a composition evaluation system.

Images