200923860 九、發明說明： 【發明所屬之技術領域】 本發明係有關於一種互動式學習系統，特別是有關於一種結合學 習歷程建議與網路資料生成教材的互動式學習系統及方法。 【先前技術】 隨著全球資訊網的快速發展，許多容易取得之各式各樣的資料皆 透過網頁縣呈現在網際網路上’再加上近年來如遠距教學之線上學 ，的盛行’所以個此廣大且豐富的㈣來建構—個課程系統，就成 為一個迫切需要研究之課題。 同步課程教學方法 學，"^^不_點之線上同步的課程教 間jfii，㈤錢枝制步課程的鮮，贿點是不受空 才处個別化的學f模式，但必須需有極大的頻寬， 才月匕有良好的教學品質，如200fi车 八w用見 1262456號，係整人串、、、目，、月21日么σ之本國專利第 結構化搜尋的、視訊直播影音節目播映，並搭配 非同步課程教學方法 李统方法係指在不同的時間及地點，只要能連上課程 f現。非同步影片是預先將課程錄影字的型態 子型態則是經由編輯者編撰教材後放置4ίί 口 200923860 擇學習的時間，但是當使用者有疑 之本國專利第ί27877〇號，=付到解荅，如2007年4月1〗曰公告 播服務端所提供的帶狀節二I將内部儲存的繪圖教學課程或數位廣 雨出至數位電視上，以供使用者學習。 混合式課程教學方法 常以同步及非_教學方法，通 年來，混合式課朗步射畔的縣教材。近 公告之本罐漸被线及研發，如_年9月21曰200923860 IX. INSTRUCTIONS: [Technical Field] The present invention relates to an interactive learning system, and more particularly to an interactive learning system and method for combining learning process suggestions and network data generation teaching materials. [Prior Art] With the rapid development of the global information network, many kinds of materials that are easy to obtain are presented on the Internet through the webpage county, plus the popularity of online learning such as distance education in recent years. This vast and rich (four) construction of a curriculum system has become an urgent need for research. Synchronous course teaching methodology, "^^不_点的线同步的课程教jfii, (5) Qianzhi step-by-step course fresh, bribery is not subject to the individualized learning f mode, but must be Great bandwidth, good monthly teaching quality, such as 200fi car eight w used to see 1262456, the whole person string,,,,,,,, 21st, the day of the national patent structure search, video live Video and audio programs, and with non-synchronous course teaching method Li Tong method refers to at different times and places, as long as the course can be connected. The non-synchronized movie is the type of the pre-synchronized video word. It is the time for the editor to edit the textbook and place 4 ί ί 200923860. However, when the user has doubts about the national patent ί27877 ,, = pay to solution荅, as in April 1, 2007, the ribbon section II provided by the announcement server will internally store the drawing teaching course or digital rain on the digital TV for users to learn. The hybrid course teaching method is often based on synchronous and non-teaching methods. Nearly announced the cans are gradually being lined up and developed, such as _ September 21
PowerPoint 此上網，就可以利用電腦來收看教學課程或訓練^程細者，、要 【發明内容】 課程資料庫 本發明係揭露-種互動式學習系統，包括：一 為記錄器，用以記錄該使用者ϊ習互動介面 成覽該課程貝枓之-她行為於該個人化學f資 2課程資料產生器根據該個人化學習資料庫所記錄之該&覽行=， 個課程資料，並輸出該下二個該 琴，程資料產生系統，包括：—課程關鍵字掏取 =用以由-雜-貝料庫中操取一課程目錄之一關鍵字； 收集1§，用以根據該關鍵字自網際網路中收集並產生至少一課程資 200923860 料’以及-縣，用⑽所收集之該課程資料加时類，並將 已分類之該課程資料存放於該課程資料庫中。 本發明另揭露—種種互動式學習方法，包括：藉由-使用者學習 j介面細-職資料是否存放於—課程資料庫中；若該課程資料 ，，於該課程倾庫中，難生—個人化賴至該使用者學習互 動，I面’記錄?使崎判互動介面職該雜㈣之—劉覽行為於 二個士化學巧料庫中；根據該個人化學習資料庫所記錄之該劉覽行 ，，自_程讀庫取得下—個課程㈣；以及輸出 料至該使用者學習互動介面。 球程貝 本發明更揭露-種課程㈣產生方法，包括：由—課程資料 ΪΪ二ίί=之根據該關鍵字自網際網路中收集並產生 i課程;=於料加以分類;以及將已分類之 於/ϋ 乂發明之目的、特徵及（或）優點能更為明顯易懂，特舉出 較佳實把例，並配合本案所附圖式，詳細說明如後。、 【實施方式】 係 本發明所述之線上課程學f及建構之 =錄，用者的學習行為’配合灰色理論來預測使用 子習課程。此—互動式學習系統與方法更彻網路收集哭 i=^ri4fpider)來收集網際網路中適合作鱗程的資源，t ，，檢索技術來增加課程分類的精準度，再配 的親念’進岭成縣雜的分類。 η眼.罔路 心请ΐί第一圖，本發明之互動式學習系統係包括：使用者互勒糸 統50、自動課程產生系統55〇、課程資料〇、 者互動系 及複數·_路觀H 1G。使时絲祕5G提供使^習9〇的0 200923860 ίϊ化端舰⑽絲介面、記紐用者_覽行為、 =料编550可以從_路中= 課程ί Π i一個自動更新課程資料的機制。更進-步而ίΐ動 網ΐ網路^料的Ιί課程資料庫5GG中娜課程資料的關鍵字來進行 ==)即是系統所設 課程資料庫：二個= 程原ϊ係會將每一門課程資料做分門別類，以利課 500中的i 為了因應課程貧料自動更新機制，課程資料庫 管理去可在雜科庫500中查無所需課程資料時， 雜目錄管理^ _來新增課程目錄。另外，^程 以判斷θ :♦丄’針對沒有搜尋到任何資料^^程目錄，管理者亦可 線至遠端伺㈣較用者學習互ί介::3 進仃妹貝料查詢及_個人化課程#料賴覽環境。…、灸 100、言=七者，動系統50係包括：使用者學習互動介面 產生器及個人化課程資料 ⑴ 課 使用者。若課程資料不存在於課程資料庫500中，則 200923860 者學習互動介面100回覆給使用者表明查益此資料。 ⑵為，並且將此 (3)個人化課程資料產生器400會依據個人化 ::料庫_中的資訊來推薦使用者適合 更進一步而言，使用者學習互動介面1〇〇是一個 動ΐ面。學f行為紀錄器·用來記錄吏用 者的概订為。個人化學習資料庫300是一 行為的資料，。個人化課程資料產生器400用來 料，以及推薦使財下—次可供學習的課 ^ 1 = 習資料庫及課程資料庫5GG的資訊配二j 匕， =二η⑼ 二來者便= 者學習互動介面100上的顯示格式有:目、=及在使用 時，則在課程資料=====在2程資料庫刪 的更新時間與課程資料最近一i争叙沾=科之則，先計鼻系統所設定 此一課程資料是否真的需要更新的，藉以判斷 包括刪除沒有内容的子目軸收j 内容， 出至使用者學習互動介面1〇0。 使用者而求的内谷，最後才輸 使用者互動系統50的動作原理和流程可進-步參考第五圖。如第 200923860 詢使用者學習互動介面剛來輸入所欲查 t Γ貝Ϊ： f驟遞）。若所查詢的課程資料確實存在於 人化課程資料，㈣lif個人化課？資料產生器棚重新整理出個 1〇〇(步驟S·、’料—，人化課程資料輸出至使用者學習互動介面 辑々驟S300)。然後，學習行為記錄器 學習互動介面⑽上的瀏覽行為，並將此；===; ^ ’個人化課程資料產生器 U庫者下—赌輯⑽料(步驟S_ ^ 程目錄管理者來更新或新增課錄漏）’讓課 明回到苐一圖，自動課程產生系續孫七4x ⑴S鍵字嫩器_從繼料庫5。。中操取課程目錄的 (2)網路課程收集器7〇〇依據課程關鍵字擷 鍵字，進行網際網路收集關 執行，因此不影響正常系統的運作。t動作曰在#景 3)課私分類器8〇〇將網路課^ 整理分類後，儲存於課程到的課程資料 ⑷當課程目錄管理者發覺課程資料不夠資料。 會自動地更新對應於新的課程目錄之m1生系統卿 200923860 更進一步而言’課程關鍵字擷 擷取課程目錄的關鍵字。 網路資料庫500中 中收集相關的課程資料。更進—半盎w〇用來從廣大的網際網路 網路收集器（Intelligent Spidei： ^貧網=程收集器彻係採用 課程的資料。網路課程收集器仫=攸、·周際網路中找尋可作為 中的網頁，並依序=程_字來搜尋網際網路 所存放的網址所對應之網頁進行^二再，序對網址仔列令 當網址仵列為空時，便結束此—回合的 °^内广的動作，最後， 用來產生子目錄及重新整理所收集σ的課 集，。課程分類器800 分類器800係先過濾、由網際網路^斤收集到的課^一步而言，課程 J式結構的分析結果來進行特 取二二依據網頁 此便完成職資料的分義作。刀韻輸出至极_庫刚中，如 自，課。程產生，統55〇的動作原理和流程可進 si_:课4呈目，管理器_會判斷課程目錄是否已4更新(、二驟 產生器4G0會計算系統所設定的更 4 ^料^化3 =PowerPoint This Internet, you can use the computer to watch the teaching courses or training courses, and [invention content] Course database The invention discloses an interactive learning system, including: a recorder for recording The user is engaged in the interactive interface to view the course - she acts in the personal chemistry 2 course data generator according to the personalized learning database recorded by the & browsing line, the course materials, and output The next two pianos, the program data generation system, includes: - course keyword capture = one keyword used in a course catalogue from the - miscellaneous - billiard library; 1 § is collected, according to the key The word collects and generates at least one course from the Internet 200923860 material 'and-county, using (10) the course information plus time class, and the classified course materials are stored in the course database. The invention further discloses that various interactive learning methods include: by using - the user learns whether the j interface fine-care data is stored in the course database; if the course material is in the course, it is difficult to be- Personalization depends on the user's learning interaction, the first side of the 'record? Make the Kawasaki interactive interface job (4) - Liu Bian acts in the two chemical library; according to the personalized learning database recorded Liu Yingxing, from the _reading library to get the next course (four); and output materials to the user learning interactive interface. The method of generating a course (4) includes: collecting and generating an i course from the Internet based on the keyword data; = classifying the material; and classifying it; The purpose, features, and/or advantages of the invention may be more clearly understood, and the preferred embodiments are described in detail with reference to the accompanying drawings herein. [Embodiment] The online course science and construction of the invention described in the present invention, the learning behavior of the user's use of the grey theory to predict the use of the sub-course. This—interactive learning system and method to collect crying i=^ri4fpider) to collect resources suitable for scales in the Internet, t, search technology to increase the accuracy of course classification, and then match the parents 'Into the classification of the county into the county. η眼.罔路心ΐΐ The first picture, the interactive learning system of the present invention includes: user mutual enthusiasm 50, automatic course production system 55 〇, course materials 〇, person interaction system and plural _ road view H 1G. When the time is 5G, the 9G is provided to make the 习9〇 0 200923860 ϊ 端 端 ( (10) silk interface, remember the user _ view behavior, = material 550 can be from _ road = course Π Π i automatically update course materials mechanism. More progress - 而 ΐ ΐ 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程 课程A course material is divided into categories, in order to facilitate the automatic update mechanism of the curriculum in the course 500, the course database management can be found in the Miscellaneous Library 500 without the required course materials, miscellaneous directory management ^ _ to add Course catalog. In addition, ^ Cheng to judge θ: ♦ 丄 'for no information found ^ ^ course directory, the manager can also line to the remote server (four) more users learn each other:: 3 into the sisters shell material query and _ Personalized course #material viewing environment. ..., moxibustion 100, words = seven, the dynamic system 50 series includes: user learning interactive interface generator and personalized course materials (1) class users. If the course material does not exist in the course database 500, then the 200923860 learn interaction interface 100 replies to the user to indicate the benefit of this information. (2) Yes, and the (3) personalized curriculum data generator 400 will recommend the user according to the information in the personalization:: __, and the user learns the interactive interface. surface. The learning behavior recorder is used to record the user's outline. The personalized learning database 300 is a behavioral material. The personalized course material generator 400 is used for materials, and the recommended courses for making money-times for learning ^ 1 = Learning database and course database 5GG information matching two j 匕, = two η (9) two come = The display format on the learning interactive interface 100 is: target, = and in use, then in the course data ===== in the 2-way database, the update time and the course materials are the most recent ones. If the data of this course set by the nose system is really needed to be updated, it is judged that the content of the sub-axis is deleted, and the user learns the interactive interface 1〇0. The inner valley of the user, and finally the operation principle and flow of the user interaction system 50 can be further referred to the fifth figure. For example, in 200923860, the user is invited to learn the interactive interface, just enter the input to check t Γ 贝Ϊ: f). If the course materials being queried do exist in the humanized course materials, (4) lif personalization class? The data generator shed is reorganized by one step (step S·, 'material), and the humanized course material is output to the user learning interactive interface (S300). Then, the learning behavior recorder learns the browsing behavior on the interactive interface (10), and this; ===; ^ 'personalized course material generator U library - gambling (10) material (step S_ ^ directory manager to update Or add new lesson leaks) 'Let the class return to the first picture, the automatic course generation system continues Sun 7 4 (1) S key word _ _ from the library 5. . (2) The online course collector 7 is based on the course keyword 撷 key to perform Internet collection and execution, so it does not affect the normal system operation. t Action 曰 in #景 3) The classifier 8 〇〇 classifies the network class ^ and stores the course materials in the course (4) When the course catalog manager finds that the course materials are insufficient. M1 Health System, which corresponds to the new course catalog, will be automatically updated. 200923860 Further, the 'Course Keywords' keyword for the course catalogue. Relevant course materials are collected in the network database 500. More-in-one-in-one w〇 is used to collect data from the vast network of Internet Explorers (Intelligent Spidei: ^Last Network = Cheng Collector). Online Course Collector 仫=攸,·周际网In the road, you can find the webpage that can be used as the webpage, and search for the webpage corresponding to the web address stored in the Internet in the order of ___, and then end the webpage when the URL is listed as empty. This - the round of the ° ^ wide action, and finally, used to generate sub-directories and rearrange the collection of collected σ. Course classifier 800 classifier 800 is the first to filter, the class collected by the Internet ^In one step, the analysis result of the J-structure of the course is used to perform the specialization of the second page. This is the definition of the completed information. The output of the knife rhyme is extremely _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The action principle and process can be entered into si_: class 4, the manager _ will judge whether the course directory has been updated 4 (the second step generator 4G0 will calculate the system set more 4 ^ ^ ^ 3 =
Si，。若課程目錄未被更新且課程 =要驟 回合的課程資料產生動作Hρ、二U叫’則結束此一 程資料需要f新昧，被課程目錄管理者更新或課 課！目錄，鍵字(步驟S1_。然後，網路課i收t器70。 S1300)。接著，所收隼㈣鞋收集课权貝料的動作（步驟 斑分類M 4會叙轉齡_ _進行重整 驟刀^驟謂）。祕，_賴_至課 200923860 明參考第-圖’自動課程產生系統咖巾的 係包括：網路搜尋器710、網址操取器72()、〇〇 =40 =路^^ 7=—個網路搜尋引擎，並 I果Si,. If the course catalog is not updated and the course = the course material to be merged, the action Hρ, the second U call', the end of the course data needs f new, updated by the course catalog manager or class! Directory, key (step S1_. Then, the network class i receives t 70. S1300). Then, the collected (four) shoes collect the right to move the bait (the step classification M 4 will be the age of _ _ to carry out the reorganization of the knife ^ sudden). Secret, _ Lai _ to class 200923860 Ming reference to the figure - the system of automatic course production system coffee towel includes: network searcher 710, web site operator 72 (), 〇〇 = 40 = road ^ ^ 7 = - Web search engine, and I fruit
力，係利用課程資料的關鍵字來對網際 P 果，若未搜尋到資料，則會停止動作並將此尋結 :之後加入-結束訊息。更進一步而t，網址摘取 π有網厂旦若未找到網址時，則會直接擷取;咖:= 頁内 在母-回合所搜尋的最後一個網址之後 另外， 具有此次搜尋的關鍵字之結束訊息。當件^，=加一則 會直接略過，直到此-結束訊息被傳至件收一結束訊息時 時才會被刪除，如此設計的目的是通知分類刀器=5類^ 850 的課程資料之收集回合已經結束，因而可= 關鍵字有關 作^網址狩列730用來提供網址存放 分類動 會先被抽出。網頁掏取器74〇用來 况’先進入的網址 :尸。是否為空。網頁掏取器仵 有網址所對應之全部網頁内容 物列’中的所 查網址佇列730不為空時，合^θ，虽網頁擷取器740檢 網路課程收集器7〇〇的動作肩採夺ώ 先’課程關鍵字擷取器6GG由課ί資料步參考第七圖。首 尋器刖開始進行網際網路搜(牛步^⑽〇)。接著，網路搜 尋器710會傳回搜尋的結果、320)。然後，網路搜 將此一搜尋結果直接記錄於課^資料廑咖f沒有找到任何網頁，則 後再進行自動更新此一課程^500中（步驟S1340)，等到以 72〇 12 200923860 ^ t 圖文内容擷取動作（步驟 f 4頁擷取器740會進行網頁的Force, using the keywords of the course materials to the Internet P results, if no data is found, it will stop the action and will find this: then join-end message. Further, t, the URL is taken π, if the network factory does not find the URL, it will directly retrieve it; the coffee: = the page after the last URL searched by the mother-round, in addition, the keyword with the search End the message. When the ^, = plus one will be skipped directly, until the end-end message is sent to the end of the message to be deleted, the purpose of this design is to inform the classification knife = 5 class ^ 850 course materials The collection round has ended, so the keyword can be used to provide the URL storage classification and will be extracted first. Web page picker 74 is used to start the URL: corpse. Whether it is empty. When the webpage crawler does not have the URL of the checked webpage 730 in the webpage content list corresponding to the webpage, the webpage picker 740 checks the action of the web course collector 7〇〇. The first step of the course keyword picker 6GG by the course ί information step reference to the seventh picture. The first searcher started the Internet search (Niubu^(10)〇). Next, the web searcher 710 will return the results of the search, 320). Then, the Internet search records this search result directly in the class ^ data 廑 f f f did not find any web page, then automatically update this course ^ 500 (step S1340), wait until 72 〇 12 200923860 ^ t map Text content capture action (step f 4 page picker 740 will perform webpage
BOO 730疋否為空（步驟幻伽）。若 °為會檢查網址仔列 .址仔列7洲中所存放的網㈣，璧^ 730不7^空的仵列，則抽出網 至步驟SHOO之間的動作 仃上述步驟S1350 址件列730為空時才會結束。·、同路貝枓收集的動作會-直持續到網 括：前置ίί 、自充550令的課程分類器800係包 ^ 840 > 85〇 h MK?i!Tw 830 ' ##iit 及統-格式。更進_步而言，前 _用來過遽雜訊 的文字連結、刪除沒有咅義的 ° 係用來清除網頁内容中 樣式分析器820用來“及分^網頁網頁内容的袼式。網頁 Ϊί分析器820係根據文件的特徵5圖進一步而言，網頁 等樣式(Style)特性，來争定綱百样 圖文表格、文字大小、色彩 式結構及樹狀網頁樣式結構i結果包括單-網頁樣 !?徵存在，而樹狀網頁樣式結構代表表網頁中只有一個 卒取器830係依序利用網頁樣式 十夕個特徵存在。特徵 步而言，特徵萃取器萃取網頁内文。更進一 同的掏取方式。舉例來說，當構的分析結果，來進行不 2構時’特徵萃取器83G會#|取網^ $結果為單-網頁樣 構的分析結果為樹狀網式糊谷’而當網頁樣式結 係依據特徵來判斷網頁内文是否符合課求特 13 200923860 當的課程肓料。分類器850用來產生子目錄，以及對所選取的課程資 ，加以分類。更進-步而言，分類n 85Q料待上述結束訊息被輸入 後’ ^會開始處理全部的特徵標題’進而產生對應的子目錄，然後， 分類器850再依據子目錄將所選取的課程資料加以分類，最後，分類 =50將分類後的課程資料輸出至課程資料庫5〇〇中，成為新的課程 ^料。背景知識_用來加強網頁樣式結制分析及網頁内容的分 ϋί二步而言’背景知識_係指網頁樣式分析器82G所使用的 * y!e'Retrieval)技術’以及分類器850所使用的類神經網BOO 730疋 is empty (step illusion). If ° is to check the URL list. The network (4) stored in the 7th continent, 璧^ 730 not 7^ empty queue, then pull out the network to the action between step SHOO, the above step S1350 address list 730 It will end when it is empty. ·, the same action as the collection of the road to the Belle - straight to the network: front ίί, self-filling 550-order course classifier 800 series package ^ 840 > 85〇h MK?i!Tw 830 ' ##iit and System-format. In addition, the previous text is used to clear the text link of the noise, and the deletion is not used. The content analyzer 820 is used to clear the content of the webpage content of the webpage. The 分析ί analyzer 820 is based on the feature 5 of the file, further, the style of the web page, etc., to contend for the outline of the graphic table, the text size, the color structure, and the tree-like style structure. There is a sign, and the tree-like webpage style structure represents that only one Stroke 830 in the table webpage sequentially uses the webpage style ten-day feature. In the feature step, the feature extractor extracts the webpage text. For example, when the analysis result is constructed, the feature extractor 83G will take the result of the analysis. The result of the analysis is a tree-like network. 'When the webpage style is based on the characteristics to determine whether the webpage text meets the course information of the class 13 200923860. The classifier 850 is used to generate sub-directories, and classify the selected course resources. Step by step, The classification n 85Q expects that after the above end message is input, '^ will start processing all feature titles' to generate corresponding subdirectories, and then the classifier 850 classifies the selected course materials according to the subdirectories, and finally, the classification = 50 Export the classified course materials to the course database 5, and become a new course material. Background knowledge _ used to strengthen the analysis of web page style analysis and the content of web pages. Refers to the * y!e'Retrieval technology used by the web page style analyzer 82G and the neural network used by the classifier 850
Networks)技術。樣式檢索(style㈣士㈤)技術是 =人__習㈣基礎所刺的祖請設計理念，可轉助課程 8GG取得高品質的文件特徵並提高分類準確度。類神經網際網 ^Neural Networks)技術主要用來產生子目錄及對課程資料進行分 課程分類器800的動作原理和流程可進一步參考 ，Networks) technology. Style search (style (four) (five)) technology is = people __ Xi (four) based on the ancestral design philosophy, can be transferred to the course 8GG to obtain high-quality document features and improve classification accuracy. Neural Network (Neural Networks) technology is mainly used to generate sub-directories and classify course materials. The principle and process of classifier 800 can be further referenced.
此一網頁内容的網頁樣式結構(步驟si_。S 頁内容的動作(步驟=)8:;根=以 式結構的萃取方式合有所不鬥。筮抱ϋ頁f式、、·。構與樹狀網頁樣 内谷中選取適當的課程資料（步驟s = 會，-回合的關鍵字之全部網頁是否都已完成 (乂驟S158G)。若未完成，則繼續等待 特:二由， ，(步驟，而後再次進行步驟s· ’ j f ===徵樣式的選擇之後，才上 分類器850'中，铁後，八f的適當課程資料輸入至 私貝抖加以刀類(步驟S1_。最後，分類後的課程資料會被 14 200923860 課程資料庫500中（步驟S1610)。 底下遂提供本發明之使用者互動系統5〇的一實施例來做進—弗 同時參閱第—圖及第二圖，在此實施例中，使用者係利^ 機^^無線上網的服務連線至伺服器端的主機。更進一步說明， ^用卞係藉蝴際網關覽n 1Q連線至舰器端的使用者學習 二〇〇「。然後’使用者輸入欲查詢課程資料的關鍵字「雜凑」。此一 夂Ϊ課際網路傳入使用者學f互動介面100，據 ^子暴程貝枓庫500進行查詢的動作。若課程資料庫5〇〇中有 Ϊ以，ΐ資料存在’例如資料結構第十章，則課程資料庫5〇〇中的 维資料ί先輸出至個人化課程資料產生器棚。'然後，個人化 粗二=料產生器400會產生個人化課程資料’並將此一個人化課程資 二兩給使用者學胃絲介面⑽。如此 罔 際網關覽器Η)來下載此一課程資料，如第九圖所示。、▲罔 =時，網際網路劉覽器10透過螢幕會顯示出此一課程目錄之 :二能夠依照自己的需求去點選。同時，學習行為記錄器200 用者在使用者學習互動介面應上的瀏覽行為，例如點選超 動作’並將此—觀行為儲存於個人化學習資料庫3GG。當使 此ί課程#料的學習後，個人化課程資料產生器侧會&據 am▲予ΐ資料庫300及課程資料庫500中的資訊來推薦使用者下一 Ά的課程貧料。 下更提供本發明之自動課程產生系統550的一實施例來做進- ^ DT^f月。5月同時參閱第一圖至第四圖’在此實施例中’使用者係使 =PDA透，無線上網的服務連線至伺服器端的主機。更進一步說明， ^用者係藉_際網路難II 1Q連線朗服器端的使用者學習互動 二。紐’侧者經由使用者學習互動介面⑽輸人欲查詢課 二貝料的關鍵字「電腦網路」，來對課程資料庫5〇〇進行查詢 右課程:貝料庫5GG中沒有此—課程資料存在，則使用者學習互動介面 15 200923860 =二目錄管理請 便會500後’課程目錄管理器900 程目錄關鍵字_取動作，以第十二圖戶^的$資^首先，進行課 關鍵字依序為「OSI模式、「TnVT；~f I斤不的弟一早為例，所擷取的 會至網際網路中進行輸路搜尋器則。網路搜尋器710 頁查詢結果的關鍵字_模式」_ 四圖=的器網頁内容，並且發現有如第十 網址加入網址件列730中並進入等&的序將第十四圖所示的 此-回合搜尋的最後一個網址。虽網址擷取器720齡 字的結束訊息於網址仵列73〇中交Γ加入一則含有此次搜尋的關鍵 網頁的内容，並將所擷取的網内$中=操取器740會擷取目前 網頁擷取器740會檢查網址仔歹貝^3；，^课程分類器_。然後， 址仵列730目前還有四個網址〇疋=空。以此實施例而言，網 網址所對應的網頁，然後對匕將第一個網址抽出並開啟此一 730 ^ 動作。 、才3停止此次的網路資料收集的 然後，網路課程收集器7〇〇 類器800。首先，前置處理哭的網頁内容會被輸出至課程分 將文字的顏色全部改為黑[、清網頁内容中的文字「[編輯]」、 的格式，修改後的網頁内容係如^予的超連結、以及統一網頁内容 器820會依據前置處理器81 ^圖所示。接著，網頁樣式分析 所輸出的網頁内容來檢查網頁樣式結 200923860 t 網=結構為嶋式結構，因 留特徵標題「歷史」及「影響」所對^中頁内特容徵=== .分類器85。部徵的選擇X作後’亦即 可以用來產生_=?==徵=「財」及「影響」 的網頁内容便可以用來產生新的課& 及「影響」所對應 據特徵標題「歷史」及「影響3 話說，分類器根 來產生新的課㈣料，並將新的課ϊί料 滿意在，若課程目錄管理者對於課程目錄不 次，當使用者想再 動)此—來’可以確保此—課程資料不會同時被啟 、予t。課程祕的產生方法係由伺服器端收集廣大的網際網路資源 17 200923860 ί使類騎作，並且可峨由管理麵管麟程目錄。 ^使用者結束此次雜浦賴縣作後，本發明 會給予制者下-次可供學狀課程㈣的輯。 、、” 法，主i储屮程學f及課程建構之互動式學習系統與方 並料胸、*—種t用獻剩際網路資源來自峽構課程資料， 習ί㈣方》1—，學習的建議之系統與方法。在本發明之互動式學 中，官理者只需針對課程目錄進行新增或更新，並不需 官ΐ課程貧料的内容，如此便可節省新增及維護課程資料所 =人力。並且’在本發明之互動式學料、統與方法中，使用者 及可供上_連線服務’就可以連線至祠服器進 仃敍貝㈣+¾，真正實現「走到哪、學到哪」的精神。 本^明提供一種線上課程學習及課程建構之互動式學習系統及方 忐，亚透過以下技術來建構出整個系統的要件： ⑴利用灰色關聯度分析(Grey Reiational Analysis)的技術， 分析使用者在本系統的瀏覽行為，讓系統能提供課程建議給 使用者’進而達成互動的目的。 ° ⑵利用網路收集器（intelligent Spider)的技術，由網際網路 中自動收集可作為課程教材之資料，同時亦解決資料是否適 合作為課程以及所有主題是否涵蓋的問題。 (3) 利用樣式檢索(StyleRetrieval)的技術，讓課程分類器取得 文件特徵的資料，並決定資料的相關性，而作為日後組織課 程目錄的依據，如此便可解決大量網路資料中會參雜許多 相關性資料的問題。 (4) 利用類神經網際網路(Neural Networks)的技術，產生課程子 目錄，並將課程内容加以分類，進而解決課程資料如何自 分類的問題。 ' 雖然本發明已以如述較佳實施例揭露如上，然其並非用以限定本 18 200923860 當可做 當以如後所附之申請專利 發明。任何熟悉此技藝者，在不脫離本發明 些許更動與潤飾，因此本發明之保護範圍應广月 ， 範圍之界定為準。 … 【圖式簡單說明】 ，-圖為本發明之互動式學習系_ 第亡圖為本發明之互動式學習系統的系統 第二圖為本發明之網路課程收集器的作圖心 發明之課程分類器的動作^ 第五圖為本發明之使用者互動系統的動 =六圖為本發明之自動課程產生系統的動=程圖。 第七圖為本發明之網路課程收集器^ Θ 匕圖為本發明之課程分_的動作程圖。 f九圖為本發明之課程資料庫所存放 的榮為本發明之使用者學習互動介面回查無課程資料 知袼式 ^一圖為本發明之課程目錄管理者所收到的查無課程資料之通 第十二圖為本發明之課程目錄管理者 第十三圖為本發明之網路搜尋器二課程目錄的格式。 拽尋所得的網頁查詢結果之部分截士。予0si杈式」進行網路 ^十四圖為本發明之網址榻取騎擷取 發明之前置處理器修改後的網頁内容。 第十六圖為本發明之特徵選擇器所選擇的網 【主要元件符號說明】 1〇網際網路瀏覽器 5〇使用者互動系統 ι〇0使用者學習互動介面 2〇0學習行為記錄器 19 200923860 300 個人化學習資料庫 400 個人化課程資料產生器 500 課程資料庫 550 自動課程產生系統 600 課程關鍵字擷取器 700 網路課程收集器 710 網路搜尋器 720 網址擷取器 730 網址佇列 740 網頁擷取器 800 課程分類器 810 前置處理器 820 網頁樣式分析 830 特徵萃取器 840 特徵選擇器 850 分類器 860 背景知識 900 課程目錄管理器 20The webpage style structure of this webpage content (step si_. S page content action (step =) 8:; root = extracting by the structure of the structure is not in conflict. 筮 ϋ f page f,, · ·. Select the appropriate course material in the tree-like webpage (step s = will, - all pages of the rounded keyword have been completed (step S158G). If not completed, continue to wait for special: two by, , (step Then, after the step s· ' jf === sign style selection is performed again, the appropriate course data of the classifier 850', the iron, and the eight f are input to the private class to be sent to the knife class (step S1_. Finally, classification The subsequent course materials will be used in 14 200923860 course database 500 (step S1610). The following provides an embodiment of the user interaction system 5 of the present invention to do the same - see also the first and second figures, In this embodiment, the user connects the service of the wireless network to the host of the server. Further, the user is connected to the player at the end of the network through the inter-channel gateway. 〇〇 ". Then the user enters the information about the course materials. The keyword "Hash". This one-time network is introduced into the user interface f interactive interface 100, according to the ^Zi storm Beibei library 500 to query the action. If there is a problem in the course database 5 ΐ ΐ ΐ ' 例如 例如 例如 例如 例如 例如 例如 例如 例如 例如 例如 例如 例如 例如 例如 例如 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十 第十The course materials will be used to download the course information (10). The Internet Explorer 10 displays the course catalogue through the screen: two can select according to their own needs. At the same time, the learning behavior recorder 200 user browsing behavior on the user interaction interface, for example Click on Super Action' and store this-view behavior in the personalized learning database 3GG. After learning the course, the personalized course data generator side will be & according to am▲ to the database 300 and Information in the course database 500 to recommend users A glimpse of the course is provided. An embodiment of the automated course production system 550 of the present invention is further provided - ^ DT ^ f月. May also refers to the first to fourth figures 'in this embodiment' The user is connected to the PDA, and the wireless Internet service is connected to the host of the server. Further, the user learns to interact with the user on the WAN server. The side learns the interactive interface (10) and wants to query the keyword "computer network" of the second class material to query the course database. The right course: there is no such thing in the 5GG in the library - the course data exists. , the user learns the interactive interface 15 200923860 = The second directory management will be 500 after the 'course directory manager 900 program directory keyword _ take action, the twelfth map of the household ^ ^ ^ first, the class keyword The preface is "OSI mode, "TnVT; ~f I don't take the case of the early morning as an example, and the network will search for the search engine in the Internet. The network searcher 710 pages query the result of the keyword_mode"_ four maps = the page content of the device, and found that the tenth URL is added to the URL list column 730 and enters the order of the & The last URL of this - round search. The end message of the 720-year-old word of the URL picker is added to the content of the key page containing the search in the URL column 73, and the selected in-network $== operator 740 will retrieve At present, the web page finder 740 will check the URL 歹 ^ ^ ^;, ^ course classifier _. Then, the address queue 730 currently has four more URLs = empty. In this embodiment, the web page corresponding to the web address is then extracted and opened for the first web address. Only 3 stops the network data collection. Then, the network course collector 7 is used. First of all, the content of the webpage that is pre-processed and cried will be output to the course to change the color of the text to black [and the format of the text "[edit]" in the content of the webpage. The content of the modified webpage is such as ^ The hyperlink and the unified web content 820 will be based on the pre-processor 81. Then, the webpage style analysis outputted the webpage content to check the webpage style knot 200923860 t net=structure is a 嶋-style structure, because the feature headings “history” and “impact” are used for the intra-page special feature===. 85. After the selection of the levy X, the content of the webpage that can be used to generate _=?== levy = "finance" and "impact" can be used to generate new lessons & and "impact" corresponding feature headings "History" and "Impact 3", the classifier roots to generate new lessons (4), and the new lesson is satisfied, if the course catalog manager is not in the course catalog, when the user wants to move again) Come 'can ensure this - the course materials will not be opened at the same time. The course secret is generated by the server side to collect the vast amount of Internet resources 17 200923860 ί make the class ride, and can be managed by the management The catalogue. ^ After the user finishes the Zai Lai County project, the invention will give the system the next-level course for the course (4). ,,,,,,,,,,,,,,,,,,, The learning system and the side of the chest, * - kind of use of the remaining network resources from the gorges curriculum materials, Xiu (four) side" 1 -, the system and method of learning recommendations. In the interactive learning of the present invention, the administrator only needs to add or update the course catalog, and does not need the content of the bureaucratic curriculum, so that the new and maintenance of the course materials can be saved. And 'in the interactive materials, systems and methods of the present invention, the user and the available service _ connection service can be connected to the 器 器 仃 ( 四 四 (4) + 3⁄4, truly realize "what to go, The spirit of learning. This book provides an interactive learning system for online course learning and course construction. Fang Ya uses the following techniques to construct the requirements of the whole system: (1) Using the technology of Grey Reiational Analysis to analyze the user's The browsing behavior of the system allows the system to provide course suggestions to the user's purpose of achieving interaction. ° (2) Using the technology of the Intelligent Spider, the Internet can automatically collect information that can be used as a course material, and also whether the information is suitable for the course and whether all topics are covered. (3) Using the style retrieval (StyleRetrieval) technology, the course classifier can obtain the data of the document characteristics and determine the relevance of the data, and as a basis for organizing the course catalog in the future, it can solve a lot of network data. Many questions about related information. (4) Using the technology of Neural Networks to generate curriculum sub-categories and classify the course content to solve the problem of how the course materials are self-classified. Although the present invention has been disclosed above in the preferred embodiments as described above, it is not intended to limit the scope of the present invention. Any person skilled in the art will be able to modify and modify the invention without departing from the invention. Therefore, the scope of protection of the present invention should be broad and the scope is defined. ... [Simple description of the diagram], - The diagram is the interactive learning system of the present invention. The second diagram of the system of the interactive learning system of the present invention is the invention of the network course collector of the present invention. The action of the course classifier ^ The fifth figure is the motion diagram of the user interaction system of the present invention. The seventh figure is the online course collector of the present invention. The figure is the action chart of the course of the present invention. f Nine diagrams are stored in the course database of the present invention. The user's learning interactive interface is back to the inventor. There is no course information knowledge. The figure is the course information obtained by the course catalog manager of the present invention. The twelfth figure of the course directory manager of the present invention is the format of the second course directory of the web crawler of the present invention. Look for some of the results of the web page query results. The network is implemented in the 0si style. The fourteenth figure is the webpage of the present invention. The sixteenth figure is the network selected by the feature selector of the present invention [main component symbol description] 1 〇 Internet browser 5 〇 user interaction system ι〇 0 user learning interactive interface 2 〇 0 learning behavior recorder 19 200923860 300 Personalized Learning Library 400 Personalized Course Data Generator 500 Course Library 550 Automatic Course Generation System 600 Course Key Picker 700 Online Course Collector 710 Web Finder 720 URL Picker 730 URL Queue 740 Web Page Picker 800 Course Classifier 810 Pre-Processor 820 Web Page Style Analysis 830 Feature Extractor 840 Feature Selector 850 Classifier 860 Background Knowledge 900 Course Catalog Manager 20