JP4934804B2 - Learning support method in e-learning - Google Patents

Description

The present invention relates to a learning support method in e-learning, which includes providing appropriate feedback in the learning process of e-learning.

With the spread of computer network technology, education using computer networks, so-called e-learning, has been actively conducted in various fields. There are various forms and methods for learning by e-learning, but the method of presenting text as a means of conveying knowledge to the learner is the most basic and widely used. One of the features of e-learning is that it provides computer-based interactivity, that is, a reaction corresponding to the learner's actions. However, learning support that gives direct feedback to text-reading activities has not been adequately performed. Therefore, the e-learning feature cannot be exhibited, and as a result, the learner has only performed a learning activity equivalent to reading a printed matter.

When reading a text (sentence), marking an important part or the like, for example, underlining is a common practice. It is known that the underlined part is well remembered, and even in e-learning, the learner has a function to leave a record such as underlining or pasting a note in the sentence. Have also been proposed (see Non-Patent Document 1). In addition, studies have been made to enhance the learning effect by emphasizing underlined important parts in advance (see Non-Patent Document 2). However, conventionally, the learner only adds information by himself or receives additional information such as underline, and no feedback is given to the information added by the learner.
Kenichi Okada, Atsushi Matsushita, IPSJ Transactions, 35 (3), 468-477 (1994) Seki Yusaku, Journal of Japan Society for Educational Technology, 21, 17-20 (1994)

The inventors underline the text of the learner and store the portion as the learner's record information, while the professor underlines the important concept of the text that is considered important according to the educational objectives. The information is stored on the computer, and the diagnosis engine compares the recorded information of the underlined part as an important concept with the recorded information of the learner. It has been proposed to feed back with a learning index (degree of agreement with correct answer) (see Non-Patent Documents 3 and 4). It has been shown that this method increases motivation for learners, allows more sophisticated or drawing activities to be seen, and improves the acquisition of learning objectives.
Yoshihiro Fukunaga, Akira Takeuchi, Mune Hirashima, Hidenobu Kunichika, Education System Information Society Research Report, 19 (3), pp. 37-42 (2004) Yoshihiro Fukunaga, Mune Hirashima, Akira Takeuchi, Society for Artificial Intelligence, Advanced Learning Science and Engineering Study Group (44th), Material for Society for Artificial Intelligence, 55-60 (2005)

However, here, the underlined part is limited to punctuation marks, and the degree of coincidence is not necessarily completely determined. In addition, the degree of coincidence given here changes depending on the number of underlining of the learner, so the appropriateness of the underlining in the current situation cannot be determined by the magnitude of the value itself, and two different underlining are not possible. From the increase or decrease in the degree of coincidence in the state, it was only possible to know which relative underline is more appropriate.

Accordingly, the problem of the present invention is that the methods of Non-Patent Documents 3 and 4 are further improved, and the character string of the learning content is devised so as to be able to determine the complete matching degree. It is to provide a method in which a learner can determine the learning situation only from the comparison of the underlined result, but the feedback can be obtained with respect to his / her learning method even with a single result.

The present inventors endeavored to further improve the methods of Non-Patent Documents 3 and 4, devised how to divide the character strings of the learning content, and correct the correct and incorrect parts with the size of the whole sentence. It was found that by introducing the concept of calculating the degree of coincidence, the appropriateness can be judged by the magnitude of the degree of coincidence. In the method of the present invention, it was confirmed that feedback for one's own learning method can be obtained with a single result.

The present invention relates to a system control method for supporting learning by feedback from a teacher to a learner in e-learning using a server and a terminal , and (1) a character string of learning content is a part of a certain unit (2) marking the important part of the character string of the learning content displayed on the terminal by the terminal operation by the teacher, and the marked character string. A step of transmitting to the server and storing the break corresponding to the model solution as break information of the model solution by the teacher , and (3) a portion considered to be important for the character string of the learning content displayed on the terminal by the terminal operation by the learner Marking is performed, and the break corresponding to the marked character string is defined as the learner's break information. Step of storing and transmitted to the server Te, (4) calculated and evaluated degree of coincidence by matching delimiter information of the model solutions of teachers and delimiter information of the learner, calculation and evaluation results of the degree of coincidence, the following This is a control method for a system that performs learning support including steps fed back to the learner on the terminal as the achievement level (A) calculated by Expression (2) .
Achievement (A) = Lc / T−Lw / (WT) (2)
(In the above formula (2), Lc is the number of “learner breaks” in an arbitrary learning unit, the number matching the “exemplary solution breaks”, and Lw is “learning” in any learning unit. Among the number of the “partitions of the person”, the number that did not match the “partition of the model solution”, T is the number of “partitions of the model solution” in an arbitrary learning unit, and W is the number of “partitions of the model solution”. Represents the number of all separators.)

In the present invention, the learning content is preferably configured for each learning unit . A learning unit is a part of learning content that is learned as a unit at a time, for example, divided into learning units, chapters, or sections. In the present invention, this is defined as a learning unit.

In the present invention, the calculation / evaluation result of the degree of coincidence may be used as an understanding index (C) calculated by the following formula (1) .
Understanding index (C) = {(Lc−Lw) / T} × 100 (%) (1)
(In the above formula (1), Lc is the number of “learner breaks” in an arbitrary learning unit, the number matching the “exemplary solution breaks”, and Lw is “learning” in any learning unit. The number that does not match the “exemplary solution delimiter” among the numbers of the “partner delimiters”, T represents the number of “exemplary solution delimiters” in an arbitrary learning unit.)

The present invention is characterized in that the coincidence calculation / evaluation result is given as an achievement level (A) calculated by the following equation (2) .
Achievement (A) = Lc / T−Lw / (WT) (2)
(In the above formula (2), Lc is the number of “learner breaks” in an arbitrary learning unit, the number matching the “exemplary solution breaks”, and Lw is “learning” in any learning unit. Among the number of the “partitions of the person”, the number that did not match the “partition of the model solution”, T is the number of “partitions of the model solution” in an arbitrary learning unit, and W is the number of “partitions of the model solution”. Represents the number of all separators.)

How to improve the learning effect in e-learning is an important issue. It is a basic action to let a learner read a sentence as a means of transmitting knowledge, and the present invention that improves the learning activity by giving appropriate feedback to the learning process is considered to have a great influence. In the present invention, first, a learner is allowed to mark, for example, underline, an important part while reading a teaching material (text) shown on a computer screen. In other words, it is expected that reading comprehension can be supported by facilitating attention to the information selected by the learning material by the learner's own marking. Next, when the learner reads an arbitrary learning unit, for example, a group of learning units, chapters or sections, according to the present invention, an evaluation value representing the appropriateness of marking is given.

Therefore, the learner obtains feedback of the evaluation value from the learning activity, and re-learns while confirming and confirming his / her own way of reading the text (how to understand). In other words, it is expected that learners will be able to improve their motivation to text, deeply understand text content, or promote understanding by repeating learning by marking text several times and receiving feedback on evaluation values several times. The Compared to conventional text creation, it takes extra time to create correct information, but teachers and educational material creators need only perform marking while reading the completed text. It seems not to be.

In the present invention, e-learning means not only a method using a computer network such as the Internet, but also a method of using a computer that is not connected to the network by distributing a CD-ROM or the like as a part of information transmission means. Is included. Further, in the present invention, the teacher includes not only a teaching professor but also a person who corresponds to a teacher such as a text creator and producer.

The method of the present invention includes (1) a step of dividing a character string of learning content into a certain unit and storing the learning content and the delimiter information in a server, and (2) the teacher displaying the above-mentioned information displayed on the terminal Marking important parts of the character string of the learning content, and storing the delimiter corresponding to the marked character string in the server as delimiter information of the model solution, (3) the learner is displayed on the terminal Marking the important part of the character string of the learning content and storing the delimiter corresponding to the marked character string in the server as delimiter information of the learner; (4) delimiter information of the teacher's model solution and A step of calculating and evaluating the degree of coincidence by collating the learner's delimiter information and feeding back the result of the calculation and evaluation to the learner. Consisting of. In the present invention, a step of feeding back the degree of coincidence of the learner and detailed marking information to the teacher side can be added to the step.

In the present invention, learning content that is a character string is input to a computer, and the character string is divided into a certain unit. As a delimiter, for example, there is a method of dividing into punctuation marks, phrase units, and semantically appropriate units. Punctuation breaks can be automated very easily with simple string processing. Phrase breaks can be automated with basic natural language processing techniques. A semantically appropriate break requires manual intervention, such as a teacher correcting the phrase unit break. In the present invention, these methods are used alone or in combination. The content and delimiter information are stored on the server. The content and the delimiter information are divided into parts to be learned as one unit at a time, for example, learning units, chapters or sections, and in the present invention, these are defined as learning units.

The teacher marks the important part of the character string in the learning content displayed on the terminal by marking the underline, changing the character display color, enclosing with a frame, etc. A mark is given to. This mark serves as a model solution. This is then stored on the server.

Next, when the learner marks a part of the learning content displayed on the terminal that is considered important (a part that the learner has determined to be important), the character string is supported. Whether or not a break is marked is checked against the break information of the model solution. If the mark is attached, the answer is correct, and if not, the error is assumed.

Based on the result, in the present invention, the degree of coincidence is calculated and evaluated, and the calculation and evaluation result is fed back to the learner. As described below, there are two methods for calculating and evaluating the degree of coincidence. It should be noted that, in the degree of matching calculation, but to calculate the percentage of the match, teachers "For example, will it can be said that the discovery by Newton." And then under drawing, students "For example, will it can be said that the discovery by Newton. "When you underline," what is the question of what percentage you agree with. In this example, semantically, the student should determine that he has the desired understanding, not simply the length ratio of the matching string portion of the marking. In this example, it is determined that the delimiter has been marked by properly demarcating in advance, for example, “/ For example / by Newton / It may be said to be discovered /”. Will be considered a match.

One is to calculate and evaluate the understanding index (C) calculated by the following formula (1) as the degree of coincidence, and to feed back the value.
Understanding index (C) = {(Lc−Lw) / T} × 100 (%) (1)
(In the above formula (1), Lc is the number of “learner breaks” in an arbitrary learning unit, the number matching the “exemplary solution breaks”, and Lw is “learning” in any learning unit. The number that does not match the “exemplary solution delimiter” among the numbers of the “partner delimiters”, T represents the number of “exemplary solution delimiters” in an arbitrary learning unit.)

The evaluation of the degree of coincidence as an understanding index given here shows the validity of the learner's overall marking, and it is an evaluation that suppresses the act of not marking or over-marking. It is set as a value. However, since the value of the degree of coincidence as an understanding index changes depending on the number of underlining of the learner, it is difficult to determine the appropriateness of underlining in the current state based on the magnitude of the value itself. However, since it is possible to know which way of underlining is more appropriate from the increase or decrease in the degree of coincidence in two different underlining states, it is meaningful depending on how it is used.

Another method is a feature of the present invention, in which the achievement level (A) calculated by the following equation (2) is calculated and evaluated as a coincidence, and the value is fed back.
Achievement (A) = Lc / T−Lw / (WT) (2)
(In the above formula (2), Lc is the number of “learner breaks” in an arbitrary learning unit, the number matching the “exemplary solution breaks”, and Lw is “learning” in any learning unit. Among the number of the “partitions of the person”, the number that did not match the “partition of the model solution”, T is the number of “partitions of the model solution” in an arbitrary learning unit, and W is the number of “partitions of the model solution”. Represents the number of all separators.)

The following characteristics are required for the achievement level, which is a numerical value indicating the appropriateness of the learner's marking (appropriateness of the judgment that it is an important part). That is, (a) it has a property of increasing when the correct part increases and decreasing when the error part increases. (B) Take a value in a fixed range regardless of the size of the learning unit (the total number of breaks included in the learning unit) and the ratio of the model solution to the whole. The reason is that when the learner sees the achievement level, the appropriateness of his / her learning can be absolutely evaluated regardless of the content. (C) The purpose is not to raise the achievement level itself, but to feed back whether the understanding of the content is appropriate depending on the degree of achievement level, and to encourage reading back. Therefore, it should not give hints for improving the achievement level, such as what percentage of the model solution is marked by the learner and the ratio of the model solution to the marked part. For example, if the ratio of the model solution to the entire content is known from the value fed back, the desired amount to be marked is known, which is inappropriate.

From the above viewpoint, the achievement level (A) satisfies the condition (a). As for the condition (b), regardless of the values of T and W, the value is 1 for all correct answers and -1 for all errors, and always satisfies this condition because it always takes a value in this range. As for the condition (c), the value when not marked (before learning) is zero, and even when marking the whole without considering anything, both are evaluated as being equivalent. So it is considered reasonable.

In the present invention, the feedback of the understanding index and the achievement level is recorded together with the learning time, and the understanding index and the achievement level are presented to the learner. The learner re-learns the content with reference to these values.

In the present invention, (3) marking by a learner and (4) calculation / evaluation of the degree of coincidence and feedback to the learner of the calculation / evaluation result are distributed on CD-ROM etc. If so, it can be executed on a single personal computer not connected to the network. In the present invention, the following evaluation can be performed by a generally known method by accumulating the results on the server.

For example, the achievement level browsing of individual learners by the teacher can be performed by providing a function for presenting the achievement levels of the individual learners to the teacher as a list. With such a function, the teacher can determine the progress of learning of each learner.

In addition, a search for learners in a specific learning situation by a teacher can be performed by providing a function for extracting learners in a specified condition using the achievement level and learning time as search keys. With this function, it is possible to find learners in a specific situation, such as learners whose achievement does not increase over time or learners whose achievement tends to decrease as the learning unit progresses.

An example of specifications when the present invention is systemized will be described. First, the text specification is configured for each learning unit. This is because it is possible for learners to know whether the learning method of each learning unit is appropriate by dividing or presenting or creating the text. There is no need to learn all the way, and only necessary unit text can be learned. On the other hand, since there are a large number of learners for the teacher, by dividing each learning unit, the entire learner can be searched for each unit, and the situation can be grasped.

For example, the system configuration is as shown in FIG. First, the teacher marks, for example, underline, the presented text on the computer, and the system stores it as correct answer data. Based on the correct answer data, the system evaluates the underlining of the learner and feeds back the evaluation result to the learner. At this time, the underlining data by the teacher and the underlining data by the learner are acquired by the same mechanism.

The learner selects a learning item displayed on the screen after login authentication at the terminal, and performs a learning activity. At that time, the learning history such as the status of learning (not implemented / under studying / completed), limited learning count, cumulative learning count, cumulative learning time, last learning date, average achievement level, etc. will be displayed to the learner. Anyway. Thereby, the learner can grasp the progress of each learning item.

When the learner underlines the text that is considered important in reading and presses the save button on the terminal screen, the underline is saved and the underline for the learning item is checked against the correct answer data. The result is fed back as an evaluation value (understanding index or achievement level) of the degree of agreement. When saved once, it is counted that it has been learned once, and saving and feedback are performed for each learning item. If the underlined portion is changed by adding / erasing the saved underline once and then saved again, the underline is re-evaluated for each learning item, and the evaluation value can be received.

The above system can be added with a learner registration function, a learning history search function, a learning material authoring function, or the like.

It is a figure which shows the structure of the system of this invention.

Claims (1)

In e-learning using a server and a terminal , there is a control method for a system that provides learning support by feedback from a teacher to a learner. (1) A character string of learning content is divided into certain parts, A step of storing learning content and delimiter information in a server; (2) marking important portions of the character string of the learning content displayed on the terminal by a terminal operation by a teacher, and delimiting corresponding to the marked character string Is transmitted to the server and stored as a delimiter information of the model solution by the teacher , (3) by marking the portion considered to be important of the character string of the learning content displayed on the terminal by the terminal operation by the learner, The delimiter corresponding to the marked character string is used as delimiter information for the learner. Step of storing and transmitted to, (4) calculated and evaluated degree of coincidence by matching delimiter information of the model solutions of teachers and delimiter information of the learner, calculation and evaluation results of the degree of coincidence, the following formula ( The control method of the system which performs the learning support which consists of the step fed back to a learner on a terminal as the achievement level (A) calculated by 2) .
Achievement (A) = Lc / T−Lw / (WT) (2)
(In the above formula (2), Lc is the number of “learner breaks” in an arbitrary learning unit, the number matching the “exemplary solution breaks”, and Lw is “learning” in any learning unit. Among the number of the “partitions of the person”, the number that did not match the “partition of the model solution”, T is the number of “partitions of the model solution” in an arbitrary learning unit, and W is the number of “partitions of the model solution”. Represents the number of all separators.)

Images