200825771 九、發明說明： 【發明所屬之技術領域】 一種建立使用者之操作習慣與興趣之模型之系統及其方法 特別是指-種分析使用者的操作行為以及操作記錄來建幻n 之操作習慣與興趣之模型之系統及其方法。 【先前技術】 鲁所謂個性化智慧服務，是指關通過對使用者錄、興趣、 心理和行為習_分析，而絲地向使用者提供其可能需要的資 訊服務。所α提供個性化智慧服務，錢要建立能比較準確= 使用者的操作特性和興趣愛好的模型。關於使用者的模型的研处 和應用，近年來已有很多進展，例如：根據上網騎字建立的ς 用者興趣模型；根據調查表（checkList)建立的使用者習慣模型； 根據點醉和瀏肋容建立的使用者行為模鮮。目前更有一些 #公司在甚至試圖建立基於電腦的操作習慣的使用者麵，以經 好地為使用者提供個性化服務。 【發明内容】 由於上杉择的朗者哺獅存在—些局限性。例如 2模型的資料來源都比較單_，再加上錯誤操作產生的資糾 直^因ί產生的伽者賴麵不鱗轉確地反應出使用; 為與興趣。Α部分的使用相麵都缺乏更钢 市卜由於使用者的興趣有可能隨著時間改變而發生變化，若 200825771 繼續維護和更新使用者的模型，一旦使用者的行為或興趣發生變 化’运些模魏失去功效。因此，本發_目的在於提供一猶 立使用者之樹_與興趣之模型之錢及其奴，透過分析使 用者的操作行紅及行祕生麟觀縣建 f與興趣之麵，並不斷的咖者的操作行為以及摔作= 為來修正已建立輯作f _興狀模型，如解決上述所提到 =相㈣法準確的反應使用者的行為與興趣或無法隨使 用者田刖之行為與興趣修正之問題。 為達上述目的，本發明可以H㈣統射法兩方成 ==系統’包括有··行為分析模組、記錄分析模“ 本發明所揭露之方法，包括有下列步驟：_㈣者操作電 子裝置產生之·行為’分析使用者之操作行為轉得對摩使用 者之操作躲；讀取使用者_電子裝置產生之縣記錄， 使用者產生讀作記錄以獲得龍者之興好；依據使用 者之興趣愛好建立制者之操作霜與興趣 之板型，禮用者之操作習慣與興趣之模型被建立後，持病以八 析產生之使用者之操作特性及使用者之興趣愛好修正使用：之； 作習慣與興趣之模型。 有之才呆 ;，有關本發日狀詳轉徵與實作，颇合圖絲實齡 細5兄明如下’其内容足以使任何熟f相關技藝者了解本發明之= 術内容並據以實施，且根據本說明書所揭露之内容及圖式技 熟習相關技藝者可輕易地理解本發明相關之目的及優點r何 200825771 【實施方式】 \下如帛1目」本發贿提之建立朗者之操作習 ===:錄分析模”。、模型 i減組彻貞纽錢財摔作雪早 裝置1〇。產生之操作行為’並分析使用者產生之操作：二 ==特性;_析模_ _取使= 包子衣置1GG產生之操作記錄，並 獲得使用者之興趣愛好；模型建立模組13〇 =3= Γ分Γ_<使用者之操作特性及記錄分析模组== 得之使用者之興趣愛好來建立使用者_ = 並於該操作«與興趣之模型被建立後，持續 t200825771 IX. Description of the invention: [Technical field of invention] A system and method for establishing a model of user's operating habits and interests, in particular, an operation habit of analyzing a user's operation behavior and operation records to construct a fantasy n Systems and methods of interest models. [Prior Art] Lu's so-called personalized smart service means that through the analysis of user records, interests, psychology and behaviors, it provides users with the information services they may need. The alpha provides a personalized smart service, and the money needs to establish a model that is more accurate = user's operational characteristics and hobbies. Regarding the research and application of the user's model, there have been many advances in recent years, such as: the user interest model established based on the Internet riding word; the user habit model established according to the questionnaire (checkList); The user behavior established by Ribong is very fresh. At present, there are some #companies in the user's face even trying to establish computer-based operating habits to provide personalized services for users. [Summary of the Invention] Because of the existence of the lions of the Uesugi selection, there are some limitations. For example, the data source of the 2 model is relatively simple, and the error caused by the error operation is corrected. The use of the Α part is lacking in the steel market. Because the user's interest may change over time, if the 200825771 continues to maintain and update the user's model, once the user's behavior or interest changes, Die Wei loses its effectiveness. Therefore, the purpose of this is to provide a model of the user's tree and the model of interest and its slaves, by analyzing the user's operations and the secrets of Linsheng County to build f and interest, and constantly The operation behavior of the café and the fall of the singer = to correct the established model of the f _ singularity, such as solving the above-mentioned = phase (four) method to accurately reflect the user's behavior and interests or unable to follow the user's 刖Behavior and interest correction issues. In order to achieve the above object, the present invention can be implemented by the H (four) system of the two methods == system 'includes · behavior analysis module, record analysis module'. The method disclosed in the present invention includes the following steps: _ (four) to operate the electronic device The behavior 'analyzes the user's operation behavior and turns to the operation of the user; reads the user's county record generated by the electronic device, and the user generates the read record to obtain the dragon's happiness; according to the user's Interests and hobbies establish the operating frost and interest plate type, and the model of the user's operating habits and interests is established, and the patient's operating characteristics and user's interests are modified to use the disease: Models of habits and interests. Some have to stay; and about the daily transformation of the hair and the actual implementation, quite close to the picture of the age of 5 brothers as follows] its content is enough to make any skilled person know the invention </ RTI> </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; The next step is as follows: "This is the practice of the establishment of the Langshi." ===: Recording the model.", Model i is reduced to the group, and the money is thrown into the snow early device. Operation generated by the user: two == characteristics; _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The operating characteristics and record analysis module == the user's hobbies to build the user _ = and after the operation «the model of interest is established, continue t
彻分析獲得之使用者之操作特性及記錄分析模組12G分== 出之使用者之興趣雯好修正該操作習慣與興趣之模型。X 型之發騎叙建域用叙_胃慣雜趣之模 t明在建立使用者的操作習慣與興趣之模型之前，首先合 ^對執行有本發明龍子裝置彻進行操作（步驟210) 操作執行電子裝置100所產生之操作行為，並分 之&作#為以轉制者賴作特 讀取使用者操作執行電子裝置100所產生之操作二，二^ 用者的操作行為產生之操作記錄以取得使用者的興趣愛Γ(3 . η 200825771 齡記錄之步㈣有絲_，意即可以 、者崎作特性，也可以綠得使聽的興趣愛好。 下L先描迷收集使用者操作執行電子裝置100所產生之摔 ====取得使用者~的步 i 221;，置來操作電子裝置1〇0時的操作行為J 盤的熟練程二t析，用:的操作特性(步驟223) ’例如對鍵 率盘時數等%U爾子裝置的時數、及執行各種程式的頻 此鸯用者在—析出的使用者的操作特性並不限於 置彻後動電子裝置100，並在啟動電子裝 利、）i㈣ 了兩小時20分中12秒（也就是8扣 ^ 1890 ^ ^ 263 竭^ ;使用者同時也操作滑鼠，則行為分析模 ㉔驟射以制制相操作特性 就是取得❹者m於取得制者的操作行為（也 術，知技 棘朗者執行電抒置啦 又印。使用者在操作電子裝置10 會將使用者賴糊她权她晴、^= 200825771 (如由β見網頁產生的cookies)的形式健存在電子裝置loo中， 其中’系統記錄權會記錄有使用者的開職與停留時間， 析後可以得到如「第4A圖」所示之操作記錄表411 ;使用者記^ 播會記錄使用者輸人的資料與喜好設定，而咖I更可錄 使用者戰的_、_的時醉，經過分析後可以得到如°「第 4Β圖」所示之網補覽記錄表421與如「第4〇圖」所示 _己錄表422 ’因此如「第2C圖」所示，本發明之記錄分析 核組120在讀取並分析上述之操作記錄（步驟231、步驟如 後，便可以得知使用者的興趣愛好（步驟234)。在本實施例中， 以對使用者戦網頁的停留時間和頻率來計算·者對特 感興趣的程度為例，由於當使用者對特定網頁感興趣時，使用者 劇覽該網頁的時間會她長，_也會財戰該师，因 以假設：m丨是咖網頁的次數，M為娜靖的總數 爾該網頁所消耗的時間，τ為_斤有網頁累積的時使 用m丨細i/T的計算是可以計算出使用者對該 趣的 度，如此本發日鞭可糾為基礎，可以 者^ 偏好程度。錢-轉概狀雜 在简之後，得知使用者對該特定遊戲有:當 經由上述兩步驟（步驟22〇、步驟23〇)之 娶得使用者的操作雜與興趣愛好，接著本發 130便可以由使財的操作特性與興較好來建立使^ 習慣與興狀_ ⑹，蝴崎 200825771 :::作 … ，相田龐大，因此可以使用如資料探勘 建立#arrng) __經__產生_，依此 建立使用者的操作習慣與興趣之模型。 (牛建立模組13Q將使用者的_慣與興趣之模型建立 用==後，便可以交域他的服務程式來使用，例如由使 知作讀與興趣之模型可以得知使用者通常在中午12點 仃特定遊戲的程式，並在上午十點至十二點經常潮 =寸认戲的相關網站時，可以得知使用者的興趣在於該特定 I之上，因此可以在上午十點至十二點時提供使用者最新的遊 12點至下午三點之間，每辭小時提示使用 座…知休息的服務。本發明在使用者對服務程式提供的服務 f的操作行為（步驟_之後，會由行為分析模組彻收集 =者的操作行為並加以分析（步驟22〇)，及由記錄分析模組 讀取使用者的操作行為產生的操作記錄並加以分析（步驟 ，使得模型建立模組130可以依據使用者的操作 _ 使用者的習慣與興趣之模型.，從而使使用者_^^^ =逐漸逼近使用者的操作特性與興趣喜好。如此，本發明即可 以解决提到之使用者的模型無法準確的反應使用者的行為與興趣 或热法隨使用者當前之行為與興趣修正的問題。 事實上’上述行為分析模、组11〇收集的使用者的操作行為與 記錄分析模組120讀取的操作特性並不能紐進行分析以做為模 200825771 建立使用者之操作習慣與興趣之模型的依據，原 栋用去於上4在使用者的操作過程中可能包含了錯誤操作’例如 =33誤等’因此在行為分析模組11◦分析使用者的操作 盯為的步驟中，必須要加入過濾錯誤操作的步驟（步驟222)，例 如^當者由_戰中某個鏈結後立刻離開，則表示使用者 點選到不魅選_，此為錯誤操作，所以需要縣掉這樣的 錯_作，同樣的，在記錄分析模組12〇分析使用者的興趣喜好 的步驟中，也需要加入過濾錯誤操作產生的操作記錄的步驟（步 驟232)，例如，當使用者由點選網頁中某個鏈結後立刻離開，表 不使用者對其·選_頁不感興趣，此為錯誤操作，所以這個 網頁所產生的alkies以及難的_記錄都應該被過濾，使得 '錯誤操料_人分析_中。其中，本發賴提之錯誤操 作並不.限於上述之點選網頁中鏈結後立刻離開之操作。 ，外，本發明包含一個情緒感測模組19〇，其用以透過特定 的器材來取得烟者的生理反應後，分析取得的生理反應以得知 使用者的心理狀_ ’並以系統時間及目前顯示之資料，將使用者 的心理狀態對應到分析產生的操作特性與興趣愛好中（步驟 240)’如「第2D圖」所示，情緒感測模組·可以利用攝影技 者表情（步驟241 )，並分析拍攝得到的表情後可以 仔知使用者的情緒，或者通過動態跟蹤使用者的眼球注視活動來 判別其對螢幕内容的感興趣程度（步驟242、步驟243)，若使用 者此時正在齡_，且面部表情經過分析後得知為開心，則情 緒感測模組190會在分析出網頁的關鍵字詞後，將使用者開心的 11 200825771 情绪對應顺用者的興趣愛好中（步驟244、步驟245)，如「第 =」之主心理狀態記錄表501所示。上述之使用攝影技術分析使 者的情緒與喜好為習知技術，但本發明並不限於此，其他如使 用手指勤感測器，也可以推測使用者的心理緊張狀態。― 雖然本發日肋前述之較佳實施觸露如上，然娃非用以限 ^發明’任何熟習相像技藝者’在不脫離本發明之精神和範圍 為之__，均屬本發明之專獅_，耻本發 準，崎範圍須視本說明書所附之申請專利範圍所界定者為 【圖式簡單說明】 之方=:係本發一使―慣舆興趣之模型 程圖 弟2B圖係本發明所提取得使用者之操作特性之詳細方法流 程圖 第2C圖係本發崎取得'使用者之興趣愛好之詳細方法流 第2D難树_提取得錢者之之詳細方法流The user's operating characteristics and the record analysis module obtained by the analysis are 12G points == The interest of the user is improved. The model of the operation habit and interest is corrected. The model of the X-type is used to describe the operation habits and interests of the user. First, the operation of the dragon device of the present invention is performed (step 210). Executing the operation behavior generated by the electronic device 100, and performing the operation record generated by the operation behavior of the electronic device 100 by the reproducer In order to obtain the user's interest in love (3. η 200825771 age record step (four) has a silk _, meaning that it can be, the characteristics of the stagnation, can also be green to listen to hobbies. L first describes the collection of user operations The execution of the electronic device 100 generates the operation of the electronic device 100. 223) 'For example, the number of hours of the key rate disk, etc., and the number of times the user who executes various programs is used is not limited to the rear-mounted electronic device 100. And in the start of electronic loading,) i (four) for two hours and 20 points in 12 (that is, 8 buckles ^ 1890 ^ ^ 263 exhausted ^; the user also operates the mouse, then the behavior analysis module 24 is fired to make the phase operation characteristic is to obtain the operator's operation behavior (also, The singer of the stalker performs the sputum and prints. The user operates the electronic device 10 and the user will be in the form of an electronic device in the form of a clear copy of the user’s rights, ^= 200825771 (such as cookies generated by the β-page). In the loo, the 'system recording right record will record the user's opening and staying time. After the analysis, the operation record table 411 as shown in "Fig. 4A" can be obtained; the user record will record the user's input. The data and preferences are set, and the coffee I can record the _, _ of the user's battle. After analysis, you can get the net browsing record table 421 as shown in the "4th drawing" and the "4th drawing". As shown in the "2C chart", the record analysis core group 120 of the present invention reads and analyzes the above operation record (step 231, after the step, the user can be known) Interests (step 234). In this embodiment, The dwell time and frequency of the page are used to calculate the degree of interest of the person. For example, when the user is interested in a particular web page, the user will be able to watch the web page for a long time, and will also fight the teacher. Because the hypothesis: m丨 is the number of times of the coffee page, M is the total amount of Najing’s time spent on the webpage, and τ is _jin. When the webpage is accumulated, the calculation can be calculated using m丨fine i/T. The degree of interest, so the current day whip can be corrected as the basis, can be ^ preference level. Money - transfer profile after the simple, know that the user has a specific game: when through the above two steps (step 22 〇, step 23〇) can get the user's operation miscellaneous and hobbies, then the hair 130 can be established by making the operation characteristics of the wealth and good to make ^ habits and developments _ (6), 蝶崎200825771 :: : for..., Xiangtian is huge, so you can use the data exploration to create #arrng) __ via __ generate _, according to which to establish the user's operating habits and interest models. (The cow building module 13Q can use the user's model of interest and interest ==, then it can be used by his service program. For example, by knowing the model of reading and interest, the user is usually At 12 o'clock in the afternoon, the program of a specific game, and from 10:00 to 12:00 in the morning, the user's interest lies in the specific I, so it can be at 10 o'clock in the morning. At 12 o'clock, the user is provided with the latest tour between 12 o'clock and 3 o'clock in the afternoon, and each hour is prompted to use the service of knowing the rest. The operation behavior of the service f provided by the user to the service program (step _ after The behavior analysis module collects the operation behavior of the person and analyzes it (step 22), and the operation record generated by the record analysis module reads the user's operation behavior and analyzes (steps, so that the model is built The group 130 can be based on the user's operation _ user's habits and interest models, so that the user _^^^ = gradually approaching the user's operating characteristics and interest preferences. Thus, the present invention can solve The user's model cannot accurately reflect the user's behavior and interests or the thermal law's problem with the user's current behavior and interest correction. In fact, the above behavioral analysis model, group 11〇 collected user's operational behavior and record The operation characteristics read by the analysis module 120 cannot be analyzed as the basis for establishing the model of the user's operating habits and interests as the model 200825771. The original building used in the above 4 may contain errors during the user's operation. The operation 'for example, '33 error, etc.', therefore, in the step of analyzing the user's operation of the behavior analysis module 11 ,, the step of filtering the error operation must be added (step 222), for example, if the person is _ in the battle If you leave immediately after the link, it means that the user clicks on the enchantment _, this is an erroneous operation, so it is necessary to make such a mistake. Similarly, in the record analysis module 12, the user's interest is analyzed. In the step, the step of filtering the operation record generated by the error operation is also required (step 232), for example, when the user leaves the link immediately after clicking a link in the webpage, the table is not used. They are not interested in the selection of the page, this is the wrong operation, so the alkies and difficult _ records generated by this page should be filtered, so that 'wrong _ _ people analysis _. Among them, this hair The erroneous operation is not limited to the operation of leaving the link immediately after the link in the above-mentioned click-selected webpage. In addition, the present invention includes an emotion sensing module 19A for obtaining a physiological response of the smoker through a specific device. Analyze the acquired physiological response to know the user's mental state _ 'and correspond to the user's psychological state to the operational characteristics and hobbies generated by the analysis (step 240) as in the system time and the currently displayed data. As shown in Fig. 2D, the emotion sensing module can use the photographic artist's expression (step 241), and analyze the captured expression to know the user's emotions, or dynamically track the user's eye gaze activity. The degree of interest in the screen content is determined (step 242, step 243). If the user is at the age _, and the facial expression is analyzed and found to be happy, the emotion sensing module 190 After analyzing the keyword words of the webpage, the user's happy 11 200825771 emotions are corresponding to the hobbyist's hobbies (steps 244 and 245), as shown in the main mental state record table 501 of "#". The above-described use of photographic techniques to analyze the emotions and preferences of the occupants is a conventional technique, but the present invention is not limited thereto, and other uses of the finger sensation sensor may also presume the state of psychological stress of the user. ― Although the preferred implementation of the foregoing is disclosed above, it is not intended to limit the invention to any skilled artisan, and is not exclusive of the spirit and scope of the present invention. Lion _, shame is issued, the scope of the Saki is subject to the scope of the patent application attached to this manual is the [simplified description of the schema] =: This is the one that makes the model of the customary interest model 2B map The detailed method flow chart of the user's operation characteristics extracted by the present invention is shown in Fig. 2C. The detailed method flow of the user's hobby is obtained by the method of the user's hobby.
第3A 12 200825771 第4A圖係本發明實施例所提之操作記錄表。 第4B圖係本發明實施例所提之網站瀏覽記錄表。 第4C圖係本發明實施例所提之網頁瀏覽記錄表。 第5圖係本發明實施例所提之心理狀態記錄表。 【主要元件符號說明】 100 電子裝置 110 行為分析模組 120 記錄分析模組 130 模型建立模組 190 情緒感測模組 301 鍵盤輸入記錄表 302 滑鼠輸入記錄表 411 操作記錄表 421 網站瀏覽記錄表 422 網頁瀏覽記錄表 501 心理狀態記錄表 步驟210使用者發生操作行為 步驟220分析使用者的操作行為以獲得使用者之操作特性 步驟221收集使用者輸入之資料 步驟222過濾錯誤操作 步驟223分析使用者輸入之資料 步驟224獲得使用者之操作特性 步驟230分析使用者的操作記錄以獲得使用者之興趣愛好 13 200825771 步驟231讀取操作記錄 步驟232過濾錯誤操作 步驟233分析操作記錄 步驟234獲得使用者之興趣愛好 步驟240分析使用者的生理活動以獲得使用者之心理狀態 並將使用者之心理狀態對應至使用者之操作特性 及興趣愛好 步驟241讀取使用者的生理反應之資料 步驟242分析使用者的生理反應 步驟243獲得使用者之心理狀態 步驟244讀取系統時間及顯示資料· 步驟245對應使用者之心理狀態與使用者之操作特性及嚴 趣愛好 “ 步驟25〇 _使用者的操作特性及興趣愛好建立操作習慣 與興趣之模型 白貝 143A 12 200825771 FIG. 4A is an operation record table provided by an embodiment of the present invention. FIG. 4B is a website browsing record table according to an embodiment of the present invention. FIG. 4C is a webpage browsing record table according to an embodiment of the present invention. Figure 5 is a mental state record table proposed by the embodiment of the present invention. [Main Component Symbol Description] 100 Electronic Device 110 Behavior Analysis Module 120 Record Analysis Module 130 Model Establishment Module 190 Emotion Sensing Module 301 Keyboard Input Record Table 302 Mouse Input Record Table 411 Operation Record Table 421 Website Browsing Record Table 422 Web browsing record table 501 Psychological state record table Step 210 User action behavior step 220 Analysis user action behavior to obtain user's operation characteristics Step 221 Collect user input data Step 222 Filter error operation step 223 Analyze the user The entered data step 224 obtains the user's operational characteristics. Step 230 analyzes the user's operational record to obtain the user's interests. 13 200825771 Step 231 Read Operation Record Step 232 Filter Error Operation Step 233 Analyze Operation Record Step 234 Obtain the User's The hobby step 240 analyzes the physiological activity of the user to obtain the psychological state of the user and maps the psychological state of the user to the user's operational characteristics and interests. Step 241 reads the physiological response data of the user. Step 242 analyzes the user. Physiological response Step 243: Obtain the psychological state of the user. Step 244: Read the system time and display the data. Step 245 corresponds to the user's psychological state and the user's operational characteristics and hobbies. Step 25: User's operational characteristics and hobbies are established. Model of operating habits and interests Baibei 14