KR100701271B1 - Online Study and Test System and Method of Foreign Language Interview - Google Patents

Abstract

The present invention relates to a system and a method for learning or evaluating specific foreign language interview writing and voice answers online, and more particularly, to allow learners to enter their personal histories and create and provide interview questions accordingly. The learner inputs a writing answer or a voice answer according to the question, and analyzes / evaluates and corrects the inputted writing or voice answer to the native speaker's score for the answer and corrected writing or corrected answer. The present invention provides an online foreign language interview learning and evaluation system that enables a learner to learn and evaluate a foreign language interview by providing a voice to a learner terminal and an interview learning and evaluation method using the system.

Foreign languages, interviews, learning and assessment systems, methods

Description

Online Study and Test System and Method of Foreign Language Interview}             

1 is a block diagram schematically showing the configuration of an online foreign language interview learning and evaluation system according to the present invention;

Figure 2 is a block diagram showing a web module constituting an online foreign language interview learning and evaluation system according to the present invention.

3 is a block diagram illustrating an interview question writing / providing module constituting an online foreign language interview learning and evaluation system according to the present invention;

4 is a block diagram illustrating a text module constituting an online foreign language interview learning and evaluation system according to the present invention;

5 is a block diagram illustrating an interview analysis / evaluation module constituting an online foreign language interview learning and evaluation system according to the present invention;

Figure 6 is a block diagram showing a TTS module constituting an online foreign language interview learning and evaluation system according to the present invention.

7 illustrates an embodiment of a multiviewer provided to a learner in an online foreign language interview learning and evaluation system in accordance with the present invention.

8 is a view sequentially showing an online foreign language interview learning and evaluation method according to the present invention.

<Description of the symbols for the main parts of the drawings>

10: learner terminal 20: web module

21: register 22: program download

23: interview question / evaluation result provider 24: interview writing input unit

25: interview voice input unit 26: personal history input unit

30: Interview question writing / providing module 30A: Complete interview question providing means

30B: Means for providing student selection examples 30C: Means for providing questions

31: Personal history storage unit 32: Interview question preparation unit for each learner

33: interview question extractor 34: learner interview question storage unit

35: Interview example writing / storage 36: Interview example providing / selection

37: interview question completion

38: complete interview question storage unit for each learner

39a: Interview text questionnaire 39b: Interview voice questionnaire

40: text module 41: text analysis unit

42: text evaluation unit 43: text correction unit

44: evaluation score / correction text transmission unit 50: interview analysis / evaluation module

50A: speech analysis means 50B: speech evaluation means

51: voice analysis unit 52: comparative analysis unit

53: data storage 54: segmental characteristic score evaluation unit

55: non-segmental characteristic score evaluation unit 56: comprehensive score evaluation unit

57: evaluation score storage unit 58: text file production unit

60: TTS module 61: personal history voice file production unit

62: text answer voice file producer 63: interview question voice file producer

64: new word pronunciation dictionary generation / provision unit 70: amount payment module

100: Interview Learning and Assessment Server

The present invention relates to a system and a method for learning or evaluating specific foreign language interview writing and voice answers online, and more particularly, to allow learners to enter their personal histories and create and provide interview questions accordingly. The learner inputs a writing answer or a voice answer according to the question, and analyzes / evaluates and corrects the inputted writing or voice answer to the native speaker's score for the answer and corrected writing or corrected answer. The present invention provides an online foreign language interview learning and evaluation system that enables a learner to learn and evaluate a foreign language interview by providing a voice to a learner terminal and an interview learning and evaluation method using the system.

Recently, the importance of foreign language, especially foreign language conversation, is increasing due to the trend of internationalization of industry and economy, and in particular, the case of adopting foreign language interviews such as English in the interview process is becoming common.

According to the increasing importance of foreign languages, many people spend a lot of time learning foreign languages, and various online and offline language learning programs have been established for this purpose.

However, in the case of Koreans, the oral structure is different from foreigners in general, and the structure of Korean language is fundamentally different from that of foreign languages. The situation is not coped.

In order to solve this problem, candidates for foreign language interviews are learning pronunciation and sentences through one-on-one lessons with foreign instructors, but this learning method has a problem that it takes a lot of money, and the education is made only at a predetermined time. As a result, people with busy daily lives such as office workers had a problem that their participation in learning was extremely limited.

Therefore, at any time, foreign language learning related to pronunciation or sentence interview for foreign languages can be carried out on their own, and a means for comparing and evaluating correct sentences and pronunciation of native speakers is required.

In order to meet this demand, an online language learning system such as a language learner equipped with a language program that can learn a foreign language using voice recognition or an internet has been developed and distributed.

The language learning apparatus as described above is based on a pronunciation comparison method using a speech signal processing technique. Specifically, the language learning apparatus recognizes a learner's pronunciation using a Hidden Markov Model (hereinafter referred to as "HMM"), and then, The method was compared with the characteristic data of the native speaker's voice. In particular, the accuracy of pronunciation was assessed using a method of comparing one element characteristic of the learner's speech with that of a native speaker, and the characteristic data of the phoneme among the segmental characteristics was used a lot. .

The native speaker's characteristic data is extracted from the trained AM (Acoustic Model) or the native speaker's voice data at the time of speech. However, the feature data extracted from the AM or native speaker's voice overlooks various characteristics of native speaker's individual vocal propensity, and it is not possible to expect correct evaluation of the pronunciation of a foreign language by using the data alone. There was a problem in that it was impossible to evaluate the strength or isochronism of a person, or even the evaluation could not be sure of its accuracy.

In addition, the pronunciation comparison programs as described above merely have a function of evaluating a learner's pronunciation and simply assigning a score, by collectively assigning the same weight to all segmental and non-segmental characteristics of utterance. It is not applicable to evaluating the characteristics of each element such as accent, stress, and vocal speed of a specific sentence or word. When learners use a learning device equipped with the program described above, the learner uses his / her pronunciation according to the native speaker's pronunciation. There was a problem that there is a limit to precise calibration.

In addition, the foreign language learning apparatus equipped with the conventional speech recognition program or engine has a problem that the learner cannot proceed with learning according to his or her level or personality because it does not specifically reflect the language use ability or speech characteristics of each learner. In addition, the conventional speech recognizers are not able to effectively distinguish even minimal pairs, which are the basic learning objectives, cannot effectively cope with noisy environments, and have low detection performance of key words or key phrases. It is not effective.

In addition, since a separate program for conducting interview learning using the language learner is not provided and the language learner is merely for conversation and pronunciation correction learning, the task of interview learning and evaluation pursued by the present invention can be performed. There was also a problem.

In addition, currently operating online English learning system is usually to pay the monthly payment and to perform the learning. Therefore, there is a problem in that learning and evaluation on a daily or weekly basis and the desired number of problems can not be carried out, which may result in financial damages for the period of not learning after monthly payment.

In order to solve the above problems, the present invention receives an individual's history of the learner online, prepares and provides an interview question for a virtual support company corresponding to the history, or inputs a personal history after the learner selects an interview question. When the learner completes the question and the learner enters an answer to the question, the learner can correct and evaluate the interview answer and provide the result to the learner, thereby fostering the learner's ability to cope with the actual foreign language interview using the online learning method. The purpose is to make it possible.

In addition, the voice input by the learner is analyzed by the voice recognition method, and compared with the data of the native speaker, and the weights for the pronunciation of foreign languages are calculated by assigning weights to factors according to segmental and non-segmental characteristics for each individual. In addition, the aim of the present invention is to provide a learner with a comparison of the interview answer and the pronunciation of the native speaker.

In addition, even when the learners input the interview answer as text, the learner himself / herself is evaluated and corrected, and the corrected text is converted into a native speaker's voice and provided to the learner with the corrected text and the evaluation score. Another aim is to cultivate the correct pronunciation of foreign languages that can be used in actual interviews by listening to and learning native speakers' voices.

In addition, by using the input feature data of the learner's voice, super-segmental voice analysis such as energy, pitch and intonation curve of the voice, spectrum analysis such as broad or narrow band spectrogram and formant basic form, and voice By analyzing and graphing variously through interpretation and time alignment, it is possible to simultaneously display at least one or more of the graphs on the display means so that the learner can check and effectively correct wrong pronunciation in a multifaceted and multifaceted manner. For other purposes.

In addition, when a new word is included in the input personal history, the speaker converts it to a native speaker's pronunciation, and adds a new word pronunciation dictionary to extract and store the voice feature data of the converted pronunciation in a database. Another aim is to make it possible to effectively conduct interview learning and evaluation through pronunciation and its characteristic data.

In addition, the learner can reduce the cost of interview learning by allowing the learner to randomly select a learning period or to select the number of questions required for interview learning and pay the amount according to the process of registering a member using a terminal. For another purpose.

In order to achieve the above object, the present invention provides a foreign language learning system connected to the learner terminal having a voice input means online, providing the interviewer according to the learner's personal history to the learner terminal, the question After receiving an answer in a voice format, the inputted voice is compared with the voice of a native speaker, and is segmented and non-segmented according to a factor-specific weight for a specific element assigned to the analyzed voice. Calculate a characteristic score and a comprehensive score, and prepare an audio file of the corrected native speaker and include an interview learning and evaluation server that provides the comparison analysis result and the score to the learner terminal. Provide a system.

In addition, using the online foreign language interview learning and evaluation system, when the learner inputs a personal history, writing an interview question using the personal history; Inputting an interview answer according to the interview question; Analyzing the input voice format answers to calculate various types of evaluation results and assigning scores by applying factor-specific weights for each individual feature element; It provides an online foreign language interview learning and evaluation method comprising the step of providing the result of the evaluation of the voice in the form desired by the learner.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings.

First, FIG. 1 illustrates an online foreign language interview learning and evaluation system according to the present invention. The learner terminal 10, which is internet-capable and has voice input / output means, is connected to the learner terminal online and asks a question necessary for an interview. Fill in the personal history and select the interview question provided or provided, and let the learner enter his or her own personal history to complete the interview question.The answer to the interview question provided or completed by the learner in text or voice form It is composed of an interview learning and evaluation server 100 for inputting, and correcting and analyzing the input interview answers and providing the results.

The interview learning and evaluation server 100 is connected to the learner terminal 10 online to allow the learner to input personal information, such as a personal history, and receive an interview answer inputted by the learner using an interview learning program or the like. A web module 20 for providing the learned results to the learners; Store the personal information such as the personal history, and when the learner requests an interview evaluation, extracts the learner's history, creates an interview question, and asks the learner through the web module 20 in text or voice form, or various types To provide learners with interview questionnaires through the web module (20), select the example questions provided by the learner, and include their own history directly in the question to complete interview questions, and convert the completed question into text or voice An interview question writing / providing module 30 to make; Extracting feature data from the answer in the form of voice to the interview question, comparing the feature data with the native speaker's feature data, and calculating the result, analyzing the segmental, non-segmental characteristics, and overall speech for the result An interview analysis / evaluation module 50 for assigning a score by applying a factor-by-factor weight, and transmitting the comparison / analysis result and the score to a web module; It consists of a TTS module 60 for converting the interview questions provided through the interview question writing / providing module 30 to a voice file.

In addition, the interview learning and evaluation server analyzes, evaluates and corrects the textual answer to the interview question, and transmits the textual module 40 for transmitting the evaluation result and the corrected text to the learner through the web module 20. The TTS module may further include a function of converting a personal history input through the web module 20 and a text form of an interview response corrected by the text module 40 into a voice file. Can be.

In addition, the interview learning and evaluation server 100 allows the learner to select the interview learning period or the number of interview questions to the learner terminal through the web module, request the payment of the amount according to the selected learning type, and the payment approved details. It may be configured to further include a payment amount module 70 to notify.

Next, as shown in FIG. 2, the web module 20 allows a learner who wants to learn and evaluate a foreign language interview to input his or her personal history, provide an interview question to the learner, and answer the answer from the learner. A register 21 for receiving an input; A program download unit 22 for allowing the learner to download a program to be basically installed for foreign language interview learning and evaluation; An interview question / evaluation result providing unit 23 for providing the learner with an interview question written according to the personal history and an evaluation result of the question; An interview voice input unit 25 for transmitting a file in a voice form input by the learner in response to the interview question to the interview analysis / evaluation module 50; It consists of a personal history input unit 26 for transmitting the learner's personal history input through the registration unit 21 to the interview question writing / providing module 30.

Here, the register 21 to select the payment type in order to transmit the necessary questions or requests and evaluation results to the connected learners to confirm whether the approval from the amount payment module 70 according to the selected payment details, and confirm Depending on the payment authorization, the interview question may be provided and an answer may be input.

In addition, the web module may further comprise a text input unit 24 for transmitting a text file of the interview response to the text module 40, and thus the interview question / evaluation result providing unit 23 May further have a function of evaluating a textual answer to the question and providing a result.

Next, the interview question writing / providing module 30 provides a completed interview question for writing and providing a completed question by using learners' personal history input through the web module 20 as shown in FIG. 3. Means 30A; Select learners to provide examples that include individual learners' personal history or examples that include specific issues, select the examples provided to modify them according to the learner's personal history, or enter specific topics of interest to complete interview questions. Example providing means 30B; And a question type providing means 30C for converting and providing the question provided from the completed interview question providing means 30A and the learner selection example providing means 30B into a question type such as text or voice.

In addition, the completed interview question providing means 30A comprises a personal history storage unit 31 for storing and managing the personal history of the learner input through the web module; When a learner having information stored in the personal history storage unit 31 is connected, the learner's interview question preparing unit for extracting the learner's personal history and using the extracted personal history to create an interview question for the learner (32); An interview question extracting unit 33 for extracting questions not provided up to that point in the interview questions written through the interview question preparing unit 32; The interview question storage unit for each learner for storing and managing the interview question extracted by the interview question extractor 33 for each learner to determine whether the question is provided later or to check the interview question provided until the learner desires It consists of 34.

In addition, the learner selection example providing means 30B includes an interview example preparation / storage unit 35 which prepares and stores an example of an interview question including a field related to a learner's individual history in advance; An interview example providing / selecting unit 36 for providing the connected learners with examples stored in the interview example creating / storing unit 35 so that the learner can select a desired example; An interview question completion unit 37 configured to complete an interview question by inputting his / her history or specific matters desired to the example selected through the interview example providing / selection unit 36; The completed interview question is configured for each learner by storing and managing the overlapping questions later, or if the learner wants to provide a complete interview question storage unit 38 for each learner to provide to view the previously completed questions.

In addition, the question type providing means 30C is an interview text questionnaire for converting the interview question transmitted from the completed interview question providing means 30A and the learner selection example providing means 30B into text to provide to the learner in text form. (34); It consists of an interview voice question unit 35 for converting the interview questions transmitted from the completed interview question providing means 30A and the learner selection example providing means 30B into a voice form.

As described above, interview questions provided for each learner can be written and provided according to the personal history of the web server or the operator, or the examples selected by the learner can be modified according to the personal history of the learner. It is very close to the types of questions you can have, which will increase your ability to adapt and learn about foreign language interviews.

Next, as illustrated in FIG. 4, the text module 40 includes: a text analyzer 41 for analyzing the answer of the learner's text form to the interview question to determine noon; The text evaluator 42 converts the interview answer in the text form of the learner into a score according to the content analyzed by the text analyzer 41, and the text form of the learner according to the content analyzed by the text analyzer 41. Text correction unit 43 for correcting the interview response of the text, the score calculated by the text evaluator 42, and the interview response corrected by the text proofing unit 43 is transmitted to the web module 20 in text form. It is composed of an evaluation score / correction text transmission unit 44 for. In addition, if the answer to the interview question does not include a voice, the interview answer in the text form is corrected and sent to the TTS module 60 to produce a voice file and provide it to the learner to practice the actual interview. It can be done.

Next, as illustrated in FIG. 5, the interview analysis / evaluation module 50 analyzes the answer in the form of voice among the interview answers transmitted from the web module 20 to extract voice and acoustic feature data. Section 51; Native speaker's feature data statistics DB 53-1 storing the feature data by analyzing the native speaker's voice, Learner's feature data statistics DB 53-2 storing the learner's feature data analyzed by the voice analyzer, The feature A data storage unit 53 composed of a theoretical value DB of the data 53-3; A PDF 52-1, Probability Density Function (DT-1), DTW (52-2), and the like are performed between the learner's feature data and the native speaker's feature data analyzed and stored in the data storage unit 53, and a specific element. Speech analysis means (50A) comprising a comparison analysis unit 52 to apply a weight (52-3) for each factor for the; The segmental characteristic score evaluator 54 calculates an evaluation score for the segmental element with respect to the interview response in the speech form analyzed by the speech analysis unit 50A, and a non-segmental basis for calculating the evaluation score for the non-segmental element. A characteristic score evaluator 55 and a comprehensive score evaluator 56 which calculates a comprehensive score for an interview answer in a voice form based on the scores calculated by the segmental and non-segmental characteristic score evaluator. And a voice evaluation means 50B configured of an evaluation score storage unit 57 for storing the scores calculated from the evaluation units according to the input voice file.

Probability Density Function (PDF) is an abbreviation of probability density function, and performing PDF between learner's feature data and native speaker's feature data means comparing the learner's feature data with native speaker's feature data. Means to indicate.

In addition, DTW (Dynamic Time Warping) is a method of recognizing speech using a template-based pattern matching method. The speech recognition system using DTW is mainly used to construct a speaker-dependent isolated word recognition system and has a high recognition rate.

In this case, in the voice analyzing means 50A, when the interview answer of the learner includes an answer in the form of a voice, the text for conversion to a text file is transmitted to the text module 40 according to the analysis of the voice analyzing unit 51. By further providing the file production unit 58, it is possible to be provided in the form of the corrected text for the portion in which the grammatical error occurred for the interview response can enhance the interview learning effect.

The characteristic data analyzed by the voice analyzer 51 refers to voice and acoustic analysis elements, and the duration of sound, energy of energy, pitch, vowels of vowels, that is, satiety Formant, Rising Falling Connection (RFC) / Tilt, Spectrum, Voice Onset Time (VOT), etc., which are divided into numerical data and pattern data. The numerical data refers to numerical data of "150Hz pitch in a certain frame", and from this, a predetermined evaluation value is calculated by performing a PDF using statistical values such as mean and standard deviation, and through the PDF The calculated estimate is for each result of performing the PDF and may be a result for a non-segmental characteristic or a result for a segmental characteristic. This is a probabilistic evaluation method, the numerical value of the data itself will have an important meaning.

In addition, the pattern data refers to the innation of the entire sentence, etc., which is not expressed as a numerical value, but rather the form, propensity, and order of the pronunciation, such as where the accent rises and falls. The back is the data, which can be compared using the DTW method. In this case, the DTW method represents a comparative evaluation method for target patterns considering temporal variation.

In addition, the learner feature data statistics DB 53-2 constituting the data storage unit 53 stores statistics of feature data of a certain number of ordinary learners divided into segmental and non-integrated items. The native speaker's feature data statistics DB 53-1 stores the statistics of the native speaker's feature data with the nearest pronunciation to the segmented and non-segmented items. In addition, the theoretical values of the characteristic data, which have been verified in the academic world or are common sense, are stored in the theoretical value DB 53-3 of the characteristic data divided into segmental and non-segmented items. In this case, the theoretical values of the feature data, for example, what kind of characteristics of the ruptured sound in general, and in the case of learners in which parts of the country or language, the error occurs a lot, such as research results and thesis so far Refers to all certified data and includes both segmental and nonsegmental feature elements.

As described above, the input data stored in the learner's voice, that is, the pronunciation data, is characterized by the native speaker's feature data statistics DB 53-1 stored in the data storage 53, the theoretical value DB 53-3 of the feature data, and the It can be said that the data obtained through a comparison result precisely through the comparison analysis unit 52.

 In addition, the comparison and analysis unit 52 performs the PDF 52-1 on the numerical data of the feature data, so that the predetermined element (sound intensity, pitch, etc.) of the learner's pronunciation is a predetermined element of the native speaker. And compare the pattern data with DTW (52-2) to compare how the learner's accent is similar to the native speaker's accent, and analyze the learner's characteristic data, the theoretical value of the feature data, and learner's utterance. Based on the data, each segmental and non-segmental tendency and importance are calculated to infer and apply the factor-weight (52-3) for each factor.

By performing PDFs on the numerical data of the learner's pronunciation and the statistics of the native speaker's characteristic data DB as described above, it is possible to know how close the numerical value is to the native speaker's value for a certain element that can be expressed as a numerical value among the learner's pronunciation. Will be. In other words, it is possible to statistically determine that certain features of a phoneme for current speech are 70% similar to native speakers.

In addition, the pattern data of the learner pronunciation extracted through the speech analyzer 51 may calculate the evaluation value by performing the DTW 52-2 with the pattern data extracted from the stored native data. As described above, the pattern data refers to intonation, tilt, and the like of segmental elements, and the DTW 52-2 compares the target patterns in consideration of temporal variation in a way to compare the pattern data. Represents comparison logic from which a value can be calculated. That is, the DTW 52-2 reinforces, normalizes, and compares linear data, which indicates how similar the pattern data to be compared is in numerical value.

In this way, by performing PDF and DTW on the characteristic data of the learner's pronunciation, that is, the numerical data and the pattern data, respectively, it is possible to generate evaluation data on the learner's utterance. The evaluation data may be divided into non-segmental characteristic evaluation data or segmental characteristic evaluation data, depending on whether the information relates to segmental elements.

Here, the segmental characteristic refers to an item that can be divided into sentences, phrases, syllables, words, phonemes, etc., and can be segmented in the case of voice, and the non-segmental characteristic is not segmented in the case of speech, such as intonation, stress, and isochronism. Say the item. That is, the peak value of the first formant of the phoneme, which is a segmental element, may be a characteristic of the segmental element, and isochronity of the input voice may be a non-segmental element.

Since the PDF and the DTW perform comparative analysis on numerical data and pattern data, each evaluation value includes both segmental and non-segmental characteristic evaluation data.

In addition, the feature evaluation data of the learner generated in this way is weighted for each element. In order to weight the learner's feature evaluation data, it is necessary to determine which segmental or non-segmental factor of the learner's pronunciation is important using the theoretical DB of the feature data of the database unit. In other words, when comparing and analyzing the data, it is possible to find out what characteristics (such as intonation, isochronism, and stress) of a learner, a phoneme, a syllable, or an utterance feature are compared with a native speaker's pronunciation. For example, if a Korean detects a characteristic such that there is little difference in pronunciation between 'r' and 'l', it is determined that the importance of the section in which such matter is detected is high. Based on the importance, weights for each non-segmented and segmented evaluation data are calculated. In other words, place weights higher on areas of high importance.

Here, the segment feature evaluation data is composed of phonemes, words, syllables, phrases, and sentences, and each segment element includes corresponding feature data and non-segmented feature evaluation data. At this time, the non-segmental characteristic evaluation data is composed of intonation, stress, negative duration, isochronous, and the like.

That is, the segmental and non-segmental characteristic evaluation data are evaluation values generated through PDF and DTW, which are weighted by element weight generation. As a result, all the features of the learner's pronunciation are calculated as a single score or a score for each segmental factor by applying the weight, which is a numerical result of the learner's pronunciation.

In addition, the speech evaluation means 50B provides analysis data of segmental or non-segmental characteristic data of the learner individual analyzed through PDF and DTW of the speech analysis means 50A and for each factor of the analysis data. According to the weights, the scores are calculated and analyzed logically by segmentation and non-segmentation of the voice input by the learner, and the scores are calculated and synthesized based on the segmental and non-segmentation scores. The analysis is performed to calculate and store the score as a numerical result. At this time, each score is stored for each corresponding voice file so that the learner can compare the results with the evaluation results inputted afterwards.

As described above, the voice interview of the learner analyzed / evaluated through the interview analysis / evaluation module 50 including the voice analysis means 50A and the voice evaluation means 50B is provided to the learners in various forms of multiviewers. The multi viewer is included in the interview question / evaluation result providing unit 23 of the web module 20. The multi viewer is a learner such as an energy curve, a pitch contour, an intonation curve, and the like. And a viewer for displaying hyperfractionated utterance analysis information among native speaker's utterance feature data; A viewer for displaying spectral analysis information among utterance characteristic data of learners and native speakers, such as wideband or narrowband spectrograms and basic forms of formants; To display utterance interpretation and time alignment information among learner and native speaker's utterance data such as phoneme level time alignment results (IPA phonetic symbols), word level time alignment results, information about the beginning and end of letters, etc. Each viewer can be displayed or simultaneously displayed on one display window.

In addition, the multi-viewer module configured as described above allows the learner to select and view the desired analysis information by clicking the mouse on the display means, and the multi-viewer display window is configured to allow the learner to select sentences or utterances that have been input so far. According to the selection, the results of the analysis are already provided in the form of a multi-viewer, and the scores are displayed so that the learner can grasp his improved records and evaluation results of various aspects at a time as he progresses the evaluation and learning. It can also induce learning motivation according to achievement. Specifically, as shown in FIG. 7, one side is provided with a plurality of function keys related to a question number, question type, my score, and voice reproduction, and the other side has a graph representing sound length, intonation, wavelength energy, and the like. Is displayed.

Next, the TTS module converts a personal history or a text file into a voice file as shown in FIG. 6, and produces a personal history voice file inputted through a web module as a voice file. Producing unit 61, the text answer voice file producing unit 62 for producing a text corrected by the text module 40 to the learner terminal to provide to the learner terminal, and included in the personal history or interview answers input through the web module A new word pronunciation dictionary generation / provider 64 for generating a new word as a native speaker's pronunciation and an interview question written in an interview question writing / providing module 30 with reference to the new word pronunciation dictionary as a voice file. It is composed of an interview question sound file production unit 63 for production. The TTS module configured as described above provides a text file by converting the text file into a voice file when the system requires a voice file.

Next, the amount payment module 70 includes a learning type selection unit that allows the learner terminal to select the interview learning period or the number of interview questions through the register 21 of the web module 20; According to the selected learning type, the payment request for the payment of the corresponding amount, and the payment settlement unit for notifying the details of the payment is configured. The amount payment module 70 configured as described above allows the learner to pay the amount of the desired learning period or the number of questions desired, so that the interview learning may be conducted only for a short time or a few interview questions, thereby reducing the cost required for the interview learning. It can be minimized.

Next, the learner learns and evaluates a foreign language interview using the online foreign language interview learning and evaluation system configured as described above, as shown in FIG. 8, the completed interview question of the interview question writing / providing module 30 is shown in FIG. 8. By providing the learner's individual interview question with reference to the inputted learner's personal history through the providing means 30A (S03), and providing the learner with the learner, the learner inputs the interview answer as a voice file (S05). The voice file is analyzed (S09). Thereafter, based on the analyzed voice file, an evaluation of the voice file's utterance is performed, and the result is calculated in various forms, and a score is assigned by applying a weight for each individual characteristic factor (S15). Thereafter, the learner completes the evaluation and learning of the foreign language interview by transmitting the evaluation result and the score to the learner through the web module 20.

In this case, the interview question is selected by providing a plurality of interview examples to the learner through the learner selection example providing means 30B of the interview question writing / providing module 30 (S03 '), and the learner himself / herself in the selected example. It may be modified to include the history or specific issues of (S03 ").

In addition, prior to providing the learner with an interview question, when the learner requests membership registration and online interview learning and evaluation online, the learner registers as a member by inputting personal information including personal history (S01) and registers the personal history. If the personal information is input, the method may further include registering as a member and downloading and installing a program for learning and evaluating online interviews (S02).

In addition, before the learner inputs an interview answer, the learner may select an interview answer type for the interview question provided (S04). You can also provide answer types by typing only in text, only in voice, and in voice and text.

When the text is included in the answer type as described above, the web module 20 determines whether the text file is included, and if the text file is included, transmits it to the text module to analyze and evaluate the interview writing text. In operation S07, when the analysis / evaluation of the text is completed, correcting an incorrect part of the text may be further included in operation S08.

As described above, when the interview answer includes a text form, the method may further include transmitting a text evaluation and correction result and an analysis evaluation result of the voice file to the learner through the web module 20.

In addition, when a learner enters a personal history or responds to an interview question during membership registration, it is determined whether a new word is not entered in the database of the interview analysis / evaluation module of the system (S10), and if a new word is entered, the TTS module. Generating a pronunciation dictionary of the new word using the new word pronunciation dictionary generation / provider of the new word and storing the generated file (S11); The native speaker pronounces the generated new word, extracts feature data from the pronounced voice file and the voice file (S13), and stores the data in the data storage unit of the interview analysis / evaluation module to analyze the voice file input by the learner. The method may further include using the step.

In addition, when the input interview answer is made up of text only, after the text proofing step (S08), the corrected text is produced as a voice file (S12) and the native speaker's voice produced from the words stored in the new word pronunciation dictionary. The method may further include transmitting the file to the learner (S16) along with the file, thereby allowing the learner to check the correct writing and pronunciation of the interview response.

In addition, when the input interview answer is composed only of a voice file, the voice file is analyzed (S09), and then the analyzed voice file is generated as a text file for analysis and evaluation through a text module. By providing the proofreading step further, the interview composition in the form of voices answered by the learner may be corrected and provided in the text form, thereby increasing the learner's interview learning effect.

While the present invention has been described with reference to the embodiments shown in the drawings, it is only for the purpose of illustrating the invention, and those skilled in the art to which the present invention pertains various modifications or equivalent implementations from the detailed description. It will be appreciated that examples are possible. Therefore, the true scope of the present invention should be determined by the technical spirit of the claims.

As described above, the online foreign language interview learning and evaluation system and method according to the present invention receives a learner's personal history online, writes an interview question for a virtual support company using the history, or provides an interview example. The learner modifies the selected example to provide a completed interview question, and if the learner answers the question, the learner can effectively learn the foreign language interview online by correcting and evaluating the interview answer. It is effective to allow learning even during idle time according to personal schedule.

In addition, it is possible to cultivate the ability to effectively respond to the foreign language interview situation in the field related to his own history by writing interview questions that correspond to the learner's personal history and use them for evaluation and learning of the interview.

In addition, when the learners input the interview answers in text, the evaluation and correction are provided in text form, and at the same time, a separate text-to-speech module is provided to compare with the native speaker's voice to enable not only pronunciation practice but also foreign language composition learning. This can double the effectiveness of interview learning.

In addition, the pronunciation input by the learner is analyzed by the voice recognition method, and the scores are calculated by comparing the data of native speakers with factor-weighted factors for the segmental and non-segmental factors of individual pronunciations. The compared data is graphed in the form of non-segmented speech analysis, spectral analysis, speech analysis, and time alignment, so that at least one or more graphs are displayed on the display means, so that the learner is accurately evaluated for his / her foreign language pronunciation. Learners can identify and effectively correct wrong parts of their pronunciation at various angles.

In addition, when a new word is included in the personal history to be input, a new word pronunciation dictionary is added to provide the learner's pronunciation of the new word to the learner so that the learner can more easily use the interview learning and evaluation system and the interview learning. And make the evaluation more systematic.

In addition, as described above, the number of interview questions can be selected, and the amount to be evaluated can be paid only for the selected number, or the amount corresponding to the desired period can be paid to alleviate the financial burden, thereby reducing the financial burden. Evaluation can be carried out.

Claims (23)

In the system that is connected to the learner terminal having a voice input means online to provide foreign language learning, A web module connected online with the learner terminal and interactively transmitting the learner terminal and interview questions and answers; Save the learner's personal history, create an interview question according to the personal history and provide it to the learner through the web module or select an interview example and enter the personal history through the learner terminal in the selected example to complete the interview question. An interview question writing / providing module; The feature data is extracted from the answer in the form of speech for the interview question and compared with the feature data of the native speaker, and the comparison result is calculated to analyze the segmental, non-segmental characteristics and comprehensive speech of the pronunciation. An interview analysis / evaluation module for assigning a score by applying a weight for each factor, and transmitting the comparison / analysis result and the score to a web module; And And a TTS module for converting the provided interview question and the corrected learner's voice response into a native speaker's voice. delete The method of claim 1, The interview learning and evaluation server, A learning type selection unit for selecting an interview learning period or an interview question number through a learner terminal; An amount payment unit for requesting payment of a corresponding amount according to the selected learning type and notifying a payment approved details; Online foreign language interview learning and evaluation system further comprising a payment module provided with. The method of claim 1, The interview question writing / providing module,  According to the learner's personal history, the interview questions can be written and provided more, or the interviewer can be selected through the learner's terminal, and the learner's terminal can be entered into the learner's terminal. An online foreign language learning and evaluation system. The method of claim 1, The web module, A register for allowing a learner who wants to learn and evaluate a foreign language to input his or her personal history, to provide a learner with an interview question, and to receive an answer from the learner; A program downloader for allowing the learner to download a program to be basically installed for learning and evaluating foreign language interviews; An interview question / evaluation result providing unit for providing learners with an interview question written according to the personal history and an evaluation result of the question; An interview voice input unit for transmitting a voice file input by the learner as an answer to the interview question to an interview analysis / evaluation module; A personal history input unit for transmitting the learner's personal history inputted through the registration unit to an interview question writing / providing module; Online foreign language interview learning and evaluation system, characterized in that consisting of. The method of claim 5, The interview question / evaluation result providing unit, A viewer for displaying hypersegmented utterance interpretation information among utterance feature data of a learner and a native speaker; A viewer that displays spectral analysis data of the learner and the native speaker by spectral analysis; A viewer for displaying utterance interpretation and time alignment information on utterance characteristic data of the learner and the native speaker; An online foreign language interview learning and evaluation system comprising a multiviewer for displaying at least one of the above on a display of a terminal. The method of claim 1, The interview question writing / providing module, Completed interview question providing means for completing and providing a completed question of learners input through the web module; A learner selection example providing means for comprehensively providing an interview example or examples including a specific matter, selecting a provided example for the learner to modify, or inputting a desired specific matter to complete an interview question; And A question type providing means for converting a question provided from the completed interview question providing means and a learner selection example providing means into a question type such as text or voice; Online foreign language interview learning and evaluation system, characterized in that consisting of. The method of claim 7, wherein The completed interview question providing means, A personal history storage unit for storing and managing a learner's personal history input through the web module; A learner interview question preparation unit for extracting a personal history of the learner individual when the learner stored in the personal history storage unit accesses and creating and providing an interview question to the corresponding learner using the extracted personal history; An interview question extracting unit extracting a question that is not provided up to that point in the interview questions written through the interview question preparing unit; And A provided question storage unit for each learner for storing and managing the interview question extracted by the provided interview question extracting unit for each learner so as to determine whether the question provided later is duplicated or to check the interview question provided until the learner desires; Online foreign language interview learning and evaluation system, characterized in that consisting of. The method of claim 7, wherein The learner selection example providing means, An interview example preparation / storage unit which prepares and stores an example of an interview question including a field related to learner's personal history in advance; An interview example providing / selecting unit for providing the connected learners with examples stored in the interview example preparation / storage unit so that the learner can select a desired example; An interview question completion unit configured to complete an interview question by inputting a learner's personal history or a desired specific matter to an example selected through the interview example providing / selecting unit; And A completed interview question storage unit for each learner for storing and managing the completed interview question for each learner and selecting a duplicate question later or providing a view to a previously selected and completed question if the learner wants to; Online foreign language interview learning and evaluation system, characterized in that consisting of. The method of claim 7, wherein The question type providing means, An interview text question unit for converting the interview question transmitted from the completed interview question providing unit and the learner selection example providing unit into text and providing the text to the learner in text form; And An interview voice question unit for converting the interview question in a text form transmitted from the completed interview question providing unit and the learner selection example providing unit into a voice to provide a voice form; Online foreign language interview learning and evaluation system, characterized in that consisting of. The method of claim 1, The interview analysis / evaluation module, The voice analysis unit extracts voice and acoustic feature data by analyzing a file in the voice format among the interview answers sent from the web module, and a native speaker feature data statistics DB for analyzing the native speaker's voice and storing the feature data; Learner feature data statistics DB for storing feature data about the pronunciation of the learner analyzed in the speech analysis unit, the data storage unit consisting of the theoretical value DB of the feature data, and the learner feature data and the native speaker voice Speech analysis means configured to perform a PDF (Probability Density Function) with the feature data, Dynamic Time Warping (DTW) and to apply a weighting factor for a specific element; And A segmental characteristic score evaluator for calculating an evaluation score for the segmental elements with respect to the interview speech answer analyzed by the speech analysis module, a non-segmental characteristic score evaluator for calculating an evaluation score for the non-segmental elements, and the segment Comprehensive score evaluation unit for calculating the overall score for the utterance input by the learner by applying the weight for each learner based on the scores calculated by the enemy and non-segment score evaluation unit, and scores calculated from the evaluation unit Voice evaluation means comprising an evaluation score storage for classifying and storing according to the input voice file; Online foreign language interview learning and evaluation system comprising a. The method of claim 11, The voice analysis means, If the interview answer of the learner includes an answer in the form of a voice, the online foreign language further comprises a text file production unit for producing a text file and sending it to the text module according to the analysis of the voice analysis unit for the voice response Interview Learning and Assessment System. The method of claim 1, The interview learning and evaluation server, After receiving more text as an answer to the interview question, the text is corrected and scores are given, and the corrected text is converted to a native speaker's voice and provided to the learner along with the corrected text. And evaluation system. The method of claim 13, The interview learning and evaluation server, A web module connected to the learner online and for interactively transmitting interview questions and answers with the learner; Save the learner's personal history, create an interview question according to the personal history and provide it to the learner through the web module or select an interview example, and let the learner enter his or her personal history in the selected example to complete the interview question. Providing interview question writing / providing module; The feature data is extracted from the answer in the form of speech for the interview question and compared with the feature data of the native speaker, and the comparison result is calculated to analyze the segmental, non-segmental characteristics and comprehensive speech of the pronunciation. An interview analysis / evaluation module for assigning a score by applying a weight for each factor, and transmitting the comparison / analysis result and the score to a web module; A text module for analyzing and evaluating a textual answer to the interview question, correcting an incorrect part, and transmitting the corrected portion to a web module; And A TTS module for converting the provided interview question and corrected learner's voice response into a native speaker's voice; Online foreign language interview learning and evaluation system comprising a. The method of claim 14, The web module, A register for allowing a learner who wants to learn and evaluate a foreign language to input his or her personal history, to provide a learner with an interview question, and to receive an answer from the learner; A program downloader for allowing the learner to download a program to be basically installed for learning and evaluating foreign language interviews; An interview question / evaluation result providing unit for providing learners with an interview question written according to the personal history and an evaluation result of the question; An interview voice input unit for transmitting a voice file input by the learner as an answer to the interview question to an interview analysis / evaluation module; A personal history input unit for transmitting the learner's personal history inputted through the registration unit to an interview question writing / providing module; And A text input unit for transmitting a text file input in response to the interview question to a writing module; Online foreign language interview learning and evaluation system further comprising. The method of claim 14, The text module, A text analyzer to determine a noon by analyzing a text response to the interview question of the learner; A text evaluator converting an interview answer in a text form of a learner into a score according to the content analyzed by the text analyzer; A text proofing unit correcting an interview answer in a text form of a learner according to the content analyzed by the text analyzing unit; And An evaluation score / correction text transmission unit for transmitting the score calculated by the text evaluation unit and the interview response corrected by the text correction unit to the web module in text form; Online foreign language interview learning and evaluation system, characterized in that consisting of. The method of claim 14, The TTS module, Personal history voice file production unit for providing a personal history input through the web module to produce a voice file; A text answer voice file production unit for producing the text corrected by the text module as a voice file and providing the text to the learner terminal; A new word pronunciation dictionary generation / provision unit for generating a new word included in a personal history or an interview answer input through the web module as a native speaker's pronunciation; And An interview question voice file production unit for producing a text file of an interview content in a text format prepared by an interview question writing / providing module with reference to the new word pronunciation dictionary; Online foreign language interview learning and evaluation system, characterized in that further comprises. By using the online foreign language interview learning and evaluation system of claim 13, Writing an interview question using the personal history when the learner inputs the personal history; Inputting an interview answer by voice or text according to the interview question; Analyzing the voice response when the input answer is a voice to calculate various types of evaluation results and assigning a score by applying a weight for each individual characteristic factor; Providing a result of evaluating the speech in a form desired by a learner; If the input answer is text, transmits it to a text module for analysis, evaluation, scoring, and correction; And And transmitting the result of the evaluation of the text and the corrected text to a learner's terminal or converting the text into a native speaker's voice. delete The method of claim 18, Filling out the interview question, And providing a plurality of interview question examples, and selecting a question that the learner wants from the question examples and completing the question to include his / her history. The method of claim 18, In the step of entering the personal history or interview answer, If a new word is not included in the database of the interview analysis / evaluation module, generating a pronunciation dictionary of the new word using a new word pronunciation dictionary generator / provider of the TTS module and storing the generated file; The native speaker pronounces the generated new word, extracts feature data from the pronounced voice file and the voice file, stores the data in the data storage of the interview analysis / evaluation module, and uses the voice file for voice analysis of the input voice file. Online foreign language interview learning and evaluation method further comprising. delete The method of claim 18, In the case where the input interview answer consists only of a voice file, analyzing the voice file, Generating the analyzed voice file into a text file; Analyzing and evaluating the generated text file through a text module, and providing a corrected interview composition.

Images