TWI615728B - smart test system and method FOR optimization education video - Google Patents

Abstract

A smart teaching film inspection system and method, which utilizes a sample sub-film to be browsed by a viewer to generate a feature record, and utilizes a detection record generated by a viewer to answer a test question, extracts a feature record as an input data, and captures a detection record. As the output data, the data exploration analysis is performed to generate the analysis model, and then the feature records generated by the viewer after browsing the new teaching film are processed through the analysis model to generate the test scores, thereby judging the merits of the new sub-film.

Description

Intelligent teaching film inspection system and intelligent teaching film inspection method

The invention relates to an intelligent teaching film inspection system and a smart teaching film inspection method, in particular to a smart teaching film inspection system and a smart teaching film inspection method by using a data exploration institute.

The design of teaching films is a very important part of distance teaching. At present, online video materials provided by manufacturers are usually edited by experienced teachers. After the recording is completed, they are placed on the website to provide students with learning.

Although recording the teaching videos through experienced teachers can greatly improve the quality of the teaching films, however, recording the teaching videos is not the same as the actual teaching, especially when recording the teaching videos without the interaction of students. Unexpectedly, the quality of the teaching film is not good. A poor quality teaching film may need to be known by the viewer's reaction or reply after a period of time on the line. This will not only achieve the expected results of the teaching film, but may also undesirably reduce the student's satisfaction.

Therefore, it is necessary to use a more advanced technology to quickly understand the problem of the teaching film before the teaching film has not been put on the shelf, or in the early stage of the shelf, that is, the information that can be viewed by the viewer.

Accordingly, it is a primary object of the present invention to provide an intelligent teaching film inspection system and an intelligent teaching film inspection method to solve the above problems.

The object of the present invention is to provide an intelligent teaching film inspection system and a smart teaching film inspection method, which can be viewed by a viewer before the teaching film is not officially disclosed, or only in the early stage of publication. Record and quickly understand the bad passages in the instructional film for the issuer or producer of the instructional film to discover the problem.

The invention relates to an intelligent teaching film inspection system for detecting a video for reviewing. The new teaching film has at least one new sub-film in time sequence, and the intelligent teaching film system comprises a film storage unit and a film factor establishment. Modules, inspection modules, exploration analysis modules, and inspection analysis modules.

The video storage unit is configured to store a plurality of video for modeling videos, and each of the sample teaching videos has at least the same sub-movie in time sequence.

The film factor creation module is configured to record at least one feature record generated by the sample sub-movie being viewed by a viewer.

The detecting module has at least one test question corresponding to the sample sub-movie, and is used for recording a test record generated by the viewer to answer the test question.

The exploration analysis module takes the feature records of the sample sub-movals as input data, and uses the detection records as output data to perform data mining analysis to optimize the video viewing model.

The inspection analysis module generates at least one feature record when the viewer browses the new teaching film, and the inspection analysis module records the feature of the new sub-film, and is processed by the analysis model to generate the test score.

In addition, the feature record of the sample sub-film or the new sub-movie may further include: the brain wave record of the viewer when the sample sub-film or the new sub-film is browsed, or the time when the sample sub-film or the new sub-film is browsed. Magnification. The time multiplier of the sample sub-movie or the new sub-movie being browsed may further include: an action taken when the sample sub-movie or the new sub-movie is browsed, or the sample sub-film or the new sub-movie is repeatedly browsed. .

As described above, the intelligent teaching film inspection system, wherein the data exploration system is an association rule analysis, a classification analysis, or a cluster analysis.

Further description of the test results, the test analysis module can compare the test score corresponding to the new child film to the preset score threshold, and can judge the new child film as a new child movie or a normal new child film.

The invention is also a smart teaching film inspection method for detecting a new teaching film, the new teaching film having at least one new sub-film in time sequence. The smart teaching video method consists of the following steps:

Step 1: storing a plurality of sample teaching videos, each of which has at least the same book in time sequence;

Step two: recording at least one feature record generated by the viewer sub-movie being browsed by the viewer;

Step 3: Record the detection record generated after the viewer answers at least one test question corresponding to the sample sub-film;

Step 4: taking the feature records of the sample sub-movals as input data, and using the detection records as output data, performing data exploration analysis to optimize the generation of the analysis model;

Step 5: The feature record corresponding to the new sub-movie generated by the viewer after browsing the new teaching film is processed by the analysis model to generate the test score.

Therefore, by using the intelligent teaching film inspection system and the intelligent teaching film inspection method provided by the present invention, the exploration and analysis module can perform data exploration analysis, that is, before the teaching film is not officially disclosed, or only in the public In the early days, viewers can browse and view the information records obtained, and quickly learn the bad passages in the teaching videos for the issuer or producer of the teaching film to discover the problem.

The advantages and spirit of the present invention will be further understood from the following detailed description of the invention.

Please refer to FIG. 1. FIG. 1 is a schematic diagram of a viewer 10 using a teaching film according to the present invention. The viewer 10 is a learner, and the terminal device 12 browses and learns the content of the teaching film. The terminal device 12 communicates with the remote server 14 through the network 16. The plurality of teaching videos are stored in the server 14 . In the middle, the remote teaching generally uses the on-demand video to complete the remote teaching.

Please refer to FIG. 2. FIG. 2 is a functional association diagram of the intelligent teaching film inspection system 30 of the present invention. The present invention relates to a Smart Learning Video Reviewing System (LVRS) 30 for detecting a video for reviewing 50, and a new teaching film 50 representing a newly published teaching film having a time sequence At least one new child film 5002, in other words, sequentially cut into a number of new sub-films 5002 according to the time sequence, and then try to find out the pros and cons of these new sub-films 5002. In this way, the teaching film can be quickly checked to find out the problematic part before the teaching film has been released.

The intelligent teaching film system includes a film storage unit 40, a film factor building module 42, a detecting module 44, a prospecting analysis module 46, and an inspection analysis module 48.

Preferably, the video storage unit 40 is built in the aforementioned server 14, and even the entire intelligent teaching film inspection system 30 can be built in the aforementioned server 14. The video storage unit 40 is configured to store a plurality of video for modelings 52. Each of the sample teaching videos 52 has at least the same book 5202 in time sequence, that is, the sample teaching film 52 is cut into many times in time series. Sample sub-film 5202.

The movie factor creation module 42 is configured to record at least one feature record 54 generated by the sample sub-film 5202 being viewed by the viewer 10. These feature records 54 are, for example, the brain magnification of the viewer 10 when the sample sub-film 5202 is viewed, or the time magnification at which the sample sub-film 5202 is viewed.

The brainwave record of the viewer 10 represents, for example, the mood reflection of the viewer 10 while browsing, and also represents the emotional response to the sample sub-film 5202, such as expectation, boredom, excitement, and the like. The time magnification of the viewer 10 browsing the sample sub-film 5202 may be proportional to the size of interest of the viewer 10 for this sample sub-film 5202, regardless of the viewing activities that occur when the sample sub-film 5202 is viewed, for example, If the operation is fast, rewind, etc., or the sample sub-film 5202 is repeatedly browsed, the time magnification will be changed. In principle, the higher the time multiplier, the more the viewer 10 is interested in the sample sub-film 5202, however, Correspondence and analysis with brainwave recordings may also be caused by the fact that this sample sub-film 5202 is unclear or too difficult.

The detection module 44 has at least one test question 56 corresponding to the sample sub-film 5202. After the viewer 10 answers the test questions 56, a test record 5602 is generated. The test record 5602 is like the test score, representing the viewer 10 pairs of the sample. The effect of the sub-film 5202 learning, the detection module 44 is used to record the detection record 5602 generated by the viewer 10 for answering the test question 56.

The exploration analysis module 46 takes the feature record 54 of the sample sub-film 5202 as the input data, and uses the detection record 5602 as the output data to perform data mining 60 analysis to optimize the generation of the analysis model (Learning Video) Viewing Models) 6002. Among them, the data mining 60 is an association rule analysis, a classification analysis, or a cluster analysis.

Subsequently, after the viewer 10a browses the new instructional film 50 through the terminal device 12, at least one feature record 54 is also generated corresponding to each new sub-film 5002. The feature records 54 are also, for example, when the sample sub-film 5202 is viewed. The brain wave record of the person 10a, or the time magnification of the sample child film 5202 being viewed. The inspection analysis module 48 processes the feature record 54 of the new sub-film 5002 through the aforementioned analysis model 6002 to generate a test score 62.

Similarly, these feature records 54 generated by the new sub-film 5002 are also, for example, the brainwave recording of the viewer 10a when the new sub-film 5002 is viewed, or the time magnification of the new sub-film 5002 being viewed.

The representativeness of the brain wave record of the viewer 10a and the time magnification of the viewer is as described above, and will not be redundant here, but in principle, the higher the time multiplier, the more the viewer 10a views the sub-film 5202. Or the new sub-film 5002 is more interesting, but with the brain wave record to correspond and analyze, the result may be that the higher the time magnification, the more interested the viewer 10a, but it may also be because this sample sub-film 5202 or new sub-film 5002 is unclear or too difficult to lead. Because of this complex relationship, it is particularly suitable for the data exploration described.

Further, the test analysis module 48 compares the value of the test score 62 corresponding to the new child film 5002 with the preset score threshold 64, and can determine that the new child film 5002 is a new child film 70 or a normal new child. Film 72. If the value from the superior to the inferior to the test score 62 is large or small, the test score 62 exceeds the score threshold of 64, and vice versa, if the value from the superior to the bad test score 62 is small and large, then It is preferred that the test score 62 is less than the score threshold of 64.

For example, please refer to FIG. 3A and FIG. 3B. FIG. 3A is a definition corresponding diagram of the operation actions described in the feature record 54 of the present invention. FIG. 3B is an example correspondence diagram of the feature record 54 of the sample sub-film 5202 or the new sub-film 5002 of the present invention in time series. In Fig. 3A, it can be seen that the current viewers 10, 10a use the teaching operations commonly used in teaching videos. For example, "PLAY" is a teaching video, "RW" is a rewinding teaching video, and "PAUSE" is a suspending teaching video..., other instructions such as This is described in Figure 3A.

The "viewer" in FIG. 3B indicates which viewer 10, 10a is browsing, and the "sub-video code" indicates the code of the sample sub-film 5202 or the new sub-film 5002 which is consecutive in time series, representing a sample sub-film. 5202 or new child film 5002. "Movie Time" indicates the time, in minutes/second, required for this sample sub-film 5202 or new sub-picture 5002 to play under normal conditions. "Browse time" is the time when the sample sub-movie 5202 or the new sub-film 5002 is actually played, in minutes/second, which may not necessarily be the same as the "movie time" because of an operation action or repeated browsing. "Time Magnification" is the quotient of "Browse Time" divided by "Video Time", that is, the "Browse Time" is the magnification of "Movie Time". Both "concentration" and "meditation value" come from brainwave recordings, and the distribution of the two values is from 0 to 100.

These feature records 54 which can be used as input data can be quantified, and the detection record 5602 as the output data can also be quantified or clarified without any doubt. Therefore, it can be used as good information for data exploration, and the exploration analysis module 46 can be used as the good information for data exploration. The feature record 54 of the sample sub-film 5202 is taken as the input data, and the detection record 5602 is used as the output data, and the data exploration 60 is analyzed to optimize the generation of the analysis model 6002. Subsequently, the inspection analysis module 48 The feature record 54 of the new sub-film 5002, processed by the aforementioned analysis model 6002, produces a test score 62 that can be used to interpret the new child film 70 or the normal new child film 72.

Please refer to FIG. 4. FIG. 4 is a flow chart of the method for verifying the intelligent teaching film of the present invention. The invention is also a smart teaching film inspection method for detecting a new teaching film 50. The new teaching film 50 has at least one new sub-film 5002 in time sequence, and the smart teaching film method comprises the following steps:

Step 1 (S01): storing a plurality of sample teaching films 52, each sample teaching film 52 having at least the same book 5202 in time sequence;

Step 2 (S02): recording at least one feature record 54 generated by the viewer sub-film 5202 being viewed by the viewer 10;

Step 3 (S03): Recording the detection record 5602 generated by the viewer 10 after answering at least one test question 56 corresponding to the sample sub-film 5202;

Step 4 (S04): taking the feature record 54 of the sample sub-film 5202 as input data, and performing the data exploration 60 analysis with the detection record 5602 as the output data to optimize the generation of the analysis model 6002;

Step 5 (S05): The feature record 54 corresponding to the new sub-film 5002 generated after the viewer 10a browses the new teaching film 50 is processed by the analysis model 6002 to generate the test score 62. Step 5 (S05) may further include the following steps: comparing the test score 62 corresponding to the new sub-film 5002 to the preset score threshold 64, it can be determined that the new sub-film 5002 is a new sub-film 70 or is a normal new Sub-film 72.

Therefore, with the intelligent teaching film inspection system 30 and the intelligent teaching film inspection method provided by the present invention, the exploration analysis module 46 performs the data exploration 60 analysis, that is, before the teaching film is not officially disclosed, or only In the early stage of the disclosure, the information record obtained by the viewer 10a can be viewed and viewed, and the bad passages in the teaching film can be quickly understood, so that the issuer or producer of the teaching film can find out the problem.

The features and spirit of the present invention will be more apparent from the detailed description of the preferred embodiments. On the contrary, the intention is to cover various modifications and equivalents within the scope of the invention as claimed.

10, 10a‧‧‧ Viewers

12‧‧‧ Terminal devices

14‧‧‧Server

16‧‧‧Network

30‧‧‧Smart Teaching Film Inspection System

40‧‧‧Video storage unit

42‧‧‧film factor building module

44‧‧‧Test module

46‧‧‧Exploration Analysis Module

48‧‧‧Inspection and Analysis Module

50‧‧‧New instructional film

5002‧‧‧New Subfilm

52‧‧‧sample teaching videos

5202‧‧‧sample film

54‧‧‧Characteristic records

56‧‧‧ test questions

5602‧‧‧Test records

60‧‧‧Information exploration

6002‧‧‧ analytical model

62‧‧‧ test scores

64‧‧‧ grade threshold

70‧‧‧New Subfilm

72‧‧‧Normal new child film

1 is a schematic diagram of a viewer using a teaching film; FIG. 2 is a functional association diagram of the intelligent teaching film inspection system of the present invention; FIG. 3A is a diagram corresponding to the definition of the operation action in the feature recording of the present invention; A sample record corresponding to a time sequence of a sample sub-film or a new sub-film; and FIG. 4 is a flow chart of the smart teaching film inspection method of the present invention.

10, 10a‧‧‧ Viewers

12‧‧‧ Terminal devices

30‧‧‧Smart Teaching Film Inspection System

40‧‧‧Video storage unit

42‧‧‧film factor building module

44‧‧‧Test module

46‧‧‧Exploration Analysis Module

48‧‧‧Inspection and Analysis Module

50‧‧‧New instructional film

5002‧‧‧New Subfilm

52‧‧‧sample teaching videos

5202‧‧‧sample film

54‧‧‧Characteristic records

56‧‧‧ test questions

5602‧‧‧Test records

60‧‧‧Information exploration

6002‧‧‧ analytical model

62‧‧‧ test scores

64‧‧‧ grade threshold

70‧‧‧New Subfilm

72‧‧‧Normal new child film

Claims (8)

A smart teaching film inspection system for detecting a video for reviewing, the new teaching film having at least one new sub-film in time sequence, the intelligent teaching film system comprising: a film storage unit for Storing a plurality of video for modeling, each sample teaching film having at least the same sub-movie in time sequence; a film factor building module for recording at least the generated sample sub-movie generated by a viewer A feature record; a test module having at least one test question corresponding to the sample sub-film, and used to record a test record generated by the viewer to answer the test question; a prospecting analysis module, taking the samples The feature record of the sub-film is used as the input data, and the detection record is used as the output data, the data mining analysis is performed to optimize the generation of an analysis model (Video Viewing Model); and the inspection analysis module is After the viewer browses the new instructional video, at least one feature record is generated, and the inspection analysis module will The feature record of the new child movie is processed by the analysis model to generate a test score, wherein the sample child film or the new child film is browsed when the viewer's brain wave record and the sample child film or the new child film is The time magnification of the browsing, the test analysis module further determines that the sample sub-film or the new sub-film is a new child according to the viewer's brain wave record and the time ratio of the sample sub-film and the new sub-picture being viewed. The movie is either a normal new sub-video. The smart teaching film inspection system according to claim 1, wherein the time ratio of the sample sub-film or the new sub-picture being viewed includes: the sample sub-film or the new sub-picture is viewed The action action that occurred, or the sample sub-movie or the new sub-movie is repeatedly viewed. The intelligent teaching film inspection system according to claim 1, wherein the data mining system is an association rule analysis, a classification analysis, or a cluster analysis. ). The intelligent teaching film inspection system according to claim 1, wherein the inspection analysis module compares the detection score corresponding to the new sub-film with a preset threshold value, and can determine the new sub-film. For a new sub-video or for a normal new sub-video. A smart teaching film inspection method for detecting a new teaching film, the new teaching film having at least one new sub-film in time sequence, the smart teaching film method comprising the following steps: Step 1: storing a plurality of sample teaching videos, Each sample teaching film has at least the same sub-movie in time sequence; step 2: recording at least one feature record generated by the viewer sub-movie being viewed by a viewer; and step 3: recording the viewer's correct answer to the sample sub-movie After at least one test test, a test record is generated; step 4: taking the feature record of the sample sub-movie as input data, and using the test record as the output data, performing data exploration analysis to optimize generation An analysis model; and step 5: causing the viewer to browse the new teaching film and generate a feature record corresponding to the new child movie, and process the analysis model to generate a test result, wherein the sample child film or the new child film The viewer's brainwave record and the sample sub-film or the new sub-film are viewed when viewed Time magnification, and determining whether the sample sub-film or the new sub-video is a new sub-video or a normal new sub-video according to the viewer's brain wave record and the time ratio of the sample sub-film and the new sub-picture being viewed. . The method for verifying a smart teaching film according to claim 5, wherein the time ratio of the sample sub-movie or the new sub-movie is included: the sample sub-movie or the new sub-movie is viewed The action action that occurred, or the sample sub-movie or the new sub-movie is repeatedly viewed. For example, the smart teaching film inspection method described in claim 5, wherein the data exploration system is a correlation rule analysis, a classification analysis, or a group analysis. For example, the smart teaching film inspection method described in claim 5, wherein the intelligent teaching film inspection method further comprises the following steps: comparing the test score corresponding to the new sub-film to a preset one-time threshold , it can be judged that the new sub-film is a new sub-video or a normal new sub-video.