TW573267B - Fuzzy-based index-relation AI method for holding ratio in bullish market - Google Patents

Fuzzy-based index-relation AI method for holding ratio in bullish market Download PDF

Info

Publication number
TW573267B
TW573267B TW91107036A TW91107036A TW573267B TW 573267 B TW573267 B TW 573267B TW 91107036 A TW91107036 A TW 91107036A TW 91107036 A TW91107036 A TW 91107036A TW 573267 B TW573267 B TW 573267B
Authority
TW
Taiwan
Prior art keywords
aforementioned
patent application
scope
item
analysis method
Prior art date
Application number
TW91107036A
Other languages
Chinese (zh)
Inventor
Chien-Ming Chen
Chih-Hung Yeh
Yen-Chih Hsu
Original Assignee
Gofaser Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Gofaser Technology Co Ltd filed Critical Gofaser Technology Co Ltd
Priority to TW91107036A priority Critical patent/TW573267B/en
Application granted granted Critical
Publication of TW573267B publication Critical patent/TW573267B/en

Links

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Description

573267 A7 B7 五、發明說明(》 [發明之技術領域] 本發明係關於一種資訊分析方法,更明確地說明,本 發明是一種在股市多頭走勢中對於股市資訊的變化進行 分析的方法,並以多項技術性指標之間的關係自動地建立 模糊邏輯規則庫,而對目前股市多頭走勢應掌握的資金持 股比例提出推論結果,使投資者藉以調節手上投資資金與 作多持股資金的比例,降低風險度與提昇投資效益。 [發明背景與習知技術] 一般而言,評選具有投資價値的股票,透過下列幾個 步驟: 1 ·先找出欲購買股票的市價,其每股稅前純益及可能 發出的股息,查明每股帳面價値及估算其未來的成長比 率。 2 ·市値與淨値的比率是多少百分點?並與前幾年做 比較。 3 ·計算該股最近幾年來的最高,最低本益比及股利, 比較目前之本益比是偏高或偏低。 4 .將以上各項數據列表比較,即可作爲判斷股票的投 資價値之依據,並且逐步追蹤觀察以尋找時機。 然而,股票市場與其說是經濟系統,不如說是「資訊 系統」,這最主要係基於:股票市場中實在有太多資訊要 分析,而非單純以經濟面切入就可量化的。另外,股票市 場的訊息包含KD、MACD、RSI、日線、週線、月線、融 資/卷、成交量-----大堆資訊,對投資者要如何處理?只 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 裝 — II 訂·!— II--- 經濟部智慧財產局員工消費合作社印製 573267 A7 B7 五、發明說明(2 ) 有不具人性之電腦系統才做得到’尤指具人工智慧(AI)之 電腦系統才做得到,這也就是爲何股票市場是資訊系統, 而非經濟系統之原因之一。 傳統上在股票市場中,對於一般個人投資者而言,投 資是非常困難的。尤其是,當投資者以有限的資金而爲了 分散投資的目的希望對一些不同的投資標的,例如:台積 電、聯電或操作期貨指數,而提出投資時,大部分投資者 會因爲對股票市場不夠了解或是缺乏對風險管理與投資 獲利進行追蹤與監測而導致資金套牢或調度不順等問題 發生。 在股票市場中,大盤指數是一項重要指標。大盤指數 的走勢是判斷股票市場走「多頭」、「空頭」或「盤整」 的主要依據之一,其中,大盤指數「多空反轉」的頭部特 徵與底部特徵往往是投資者最希望能夠掌握的資訊之 一;如果能夠掌握大盤指數的頭部特徵,則投資者便能夠 風險控管,事先獲利了結避免套牢;反之,如果能夠掌握 大盤指數的底部特徵,則投資者便能夠有效調度資金進場 選股。所以在習知技術中有許多理論被用來判斷大盤走 勢,例如,波浪理論。573267 A7 B7 V. Description of the Invention ("Technical Field of the Invention") The present invention relates to an information analysis method. More specifically, the present invention is a method for analyzing changes in stock market information in the long trend of the stock market. The relationship between multiple technical indicators automatically establishes a fuzzy logic rule base, and proposes inference results on the proportion of capital holdings that should be grasped by the current bull trend in the stock market, so that investors can adjust the proportion of investment funds in hand and long holding funds [Background and know-how] Generally speaking, the selection of stocks with investment price, through the following steps: 1 · First find out the market price of the stock to be purchased, its pre-tax per share Net profit and possible dividends, find out the book price per share and estimate its future growth ratio. 2 · What percentage of the market to net ratio? And compare with the previous years. 3 · Calculate the stock in recent years The highest and lowest P / E ratios and dividends are higher or lower than the current P / E ratio. 4. Compare the above data lists, It can be used as a basis for judging the investment price of stocks, and gradually follow the observations to find the opportunity. However, the stock market is not so much an economic system as an "information system", which is mainly based on the fact that there is too much information in the stock market It should be analyzed, not quantified by simply cutting in from the economic side. In addition, the stock market information includes KD, MACD, RSI, daily, weekly, monthly, financing / volume, trading volume ----- large pile Information, how to deal with investors? Only the size of this paper applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) (Please read the precautions on the back before filling this page). — — — — — — II --- Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 573267 A7 B7 V. Description of the Invention (2) Only a computer system with no human nature can be done ', especially a computer system with artificial intelligence (AI), which can also be done. This is one of the reasons why the stock market is an information system rather than an economic system. Traditionally in the stock market, it has been very difficult for the average individual investor to invest. Especially when investing For the purpose of diversifying investment with limited funds, investors hope to invest in different investment targets, such as: TSMC, UMC, or operating futures indexes. When making an investment, most investors will lack the understanding of the stock market or lack of Risk management and investment profit tracking and monitoring lead to problems such as capital seizure or irregular scheduling. In the stock market, the broad market index is an important indicator. The trend of the broad market index is to judge the stock market to go "long" and "short". Or one of the main reasons for "consolidation". Among them, the head and bottom characteristics of the "long-short reversal" of the broad market index are often one of the most important information investors want to know. If the head characteristics of the broad market index can be grasped, Investors will be able to manage risks and gain profits in advance to avoid lock-in. On the contrary, if they can grasp the bottom characteristics of the broader market index, investors can effectively dispatch funds to enter the market and select stocks. Therefore, there are many theories in the conventional technology used to judge the broad market trend, for example, the wave theory.

波浪理論爲美國投資專家RALPH NELSON ELLIOTT所發明,是一種複雜而較難了解的技術性分析 方法,要達到較高的準確性,就必須完整的了解波浪理論 的所有原則,如果投資者僅是了解基本的皮毛,反而會做 出錯誤的分析。一般人對於波浪理論的了解大部分是限於 (請先閱讀背面之注意事項再填寫本頁)The wave theory was invented by the American investment expert RALPH NELSON ELLIOTT. It is a complex and difficult to understand technical analysis method. To achieve high accuracy, you must fully understand all the principles of the wave theory. If investors only understand The basic fur will make the wrong analysis. Most people's understanding of wave theory is limited (please read the precautions on the back before filling this page)

經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 Α7 Β7 五、發明說明(3 ) 五個主波段及三個修正波段所組成,但是,波浪理論有將 近3 0 0條原則之多,如果不是專業的分析人員,要懂得 3 0 0條的理論,實在是一件非常困難的事。 另外’在多頭巾場進入尾聲時,大盤指數雖創新高 點,但成交平均量減少時,股價隨時可能反轉,投資者應 將資金有效轉移至作空的股票,以作避險之準備;往往在 空頭市場進入尾聲時,大盤指數雖創新低點,但近期成交 平均量穩定甚至朝上,股價可能會回升,投資者應掌握可 週轉資金,目標轉移至作多的股票。成交量是股市的動 能,能源充足股價才會上漲。然而,投資者每每在股市多 頭或空頭走勢中,卻沒有對其資金與持股間的掌控作出即 時的決策,而會在多頭市場進入尾聲還滿檔持股,或空頭 市場進入尾聲卻沒有足夠資金購買起漲的股票。因此,投 資者若能對投資資金與持股資金的比例作有效的掌控,將 可降低投資風險度。 此外,影響投資者的判斷還包含經常是出現在股市之 中的消息面,以及投資者的情緒問題,這使得投資者的經 驗無法對其判斷產生有效的助益,造成投資決策錯誤的可 能性大爲提昇。因此,對於一直以來以「人」爲決策中心 的股市投資問題,發展出以「人的智慧」爲基礎的人工智 慧系統,將可從系統所接收的股市資訊實際反應推論的結 果,而不受消息面與情緒影響,達到以「人智」代替「人 治」的目的。 目前具人工智慧判斷之系統或專家系統(Expert 本紙張尺度適用中國國家標準(CNS)A4規格(2〗〇χ 297公釐) " (請先閱讀背面之注意事項再填寫本頁) 0 -線· 經濟部智慧財產局員工消費合作社印製 573267 A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明說明(4 ) System)中,「人的智慧」或「人的經驗」通常被「表示」 在一知識資料庫(Knowledge Database)或規則資料庫 (Rule-based Database),這些資料庫所建立的知識或規則係 對系統的應用領域中,輸入資料或訊息變化加以規範與監 測以產生推論(Inference)。所以,系統必須包含資料認知 (Data Acquaintance)的機制。例如,在模糊邏輯控制系統 (Fuzzy Logic Control System)中,模糊邏輯規則對於其語意 變數(Linguistic Variables)的認知係藉由歸屬函數 (Membership Function)來界定。因此,發展人工智慧判斷 股票資訊處理系統必須發展對股市資訊認知的監測機制。 在利用模糊控制方法的發展中,美國發明專利案第 5,189,728 號「Rule generating and verifying apparatus for fuzzy control」揭露一種產生規則的方法,該方法能夠對於 輸出値與輸入値使用關於其輸出與輸入語意變數的預定 歸屬函數產生有用的規則(IF-THEN rules),並且根據產生 的規則做出模糊推論(Fuzzy Inference)。 因此,本發明的主要目的係提供一種分析在股市多頭 走勢行情中,以多項技術指標之間的關係自動地組織出模 糊邏輯規則庫,推論出多頭走勢的資金持股比例之方法。 本發明的另一目的係提供一種分析在股市多頭走勢 行情中,可藉由不定期的重新組織模糊邏輯規則庫,對當 前股市多頭走勢實施適應性分析資金比例的AI方法,以 作爲人工智慧系統判斷與處理之依據。 [發明槪述] (請先閱讀背面之注意事項再填寫本頁) ·裝 ·Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, the paper size is applicable to the Chinese National Standard (CNS) A4 (210 X 297 mm) 573267 Α7 Β7 V. Description of the invention (3) Five main bands and three modified bands However, there are nearly 300 principles in the wave theory. If you are not a professional analyst, it is very difficult to understand the 300 theories. In addition, at the end of the turban market, although the broad market index reached a new high, but when the average transaction volume decreased, the stock price may reverse at any time. Investors should effectively transfer funds to short stocks in preparation for risk aversion; Often when the short market comes to an end, although the broad market index hits a new low, the average transaction volume has recently stabilized or even turned upwards, and the stock price may rebound. Investors should have working capital and transfer their targets to long stocks. The trading volume is the momentum of the stock market, and the energy sufficient stock price will rise. However, investors often do not make an immediate decision on the control between their funds and holdings when the stock market is in a bullish or bearish trend, and they will enter the end of the bullish market and have full-scale holdings, or the bearish market will not end. Money to buy rising stocks. Therefore, if investors can effectively control the ratio of investment funds to shareholding funds, the risk of investment will be reduced. In addition, the judgments affecting investors also include news faces that often appear in the stock market, as well as investor sentiment problems, which makes the experience of investors unable to effectively help their judgments and the possibility of making wrong investment decisions. Greatly improved. Therefore, for the stock market investment problem that has always been based on "people" as the decision center, an artificial intelligence system based on "human wisdom" has been developed. The influence of news and emotions achieves the goal of replacing "rule of man" with "human intelligence". Currently a system or expert system with artificial intelligence judgment (Expert This paper size applies to China National Standard (CNS) A4 specification (2) 0 × 297 mm) " (Please read the precautions on the back before filling this page) 0- Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 573267 A7 B7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. Inventory (4) System), "Human Wisdom" or "Human Experience" is usually " "Representation" is in a Knowledge Database or Rule-based Database. The knowledge or rules established in these databases are to regulate and monitor the changes in input data or information in the application field of the system. Inference. Therefore, the system must include a mechanism for data cognition (Data Acquaintance). For example, in the Fuzzy Logic Control System, the recognition of Linguistic Variables by fuzzy logic rules is defined by the Membership Function. Therefore, the development of artificial intelligence judgment stock information processing system must develop a monitoring mechanism for stock market information cognition. In the development of the use of fuzzy control methods, U.S. Patent No. 5,189,728 "Rule generating and verifying apparatus for fuzzy control" discloses a method of generating rules that can be used for output 値 and input 关于 with respect to its output and input. Predetermined assignment functions of semantic variables generate useful rules (IF-THEN rules) and make fuzzy inferences based on the generated rules. Therefore, the main object of the present invention is to provide a method for automatically analyzing the fuzzy logic rule base based on the relationship between multiple technical indicators in the stock market bullish trend market, and inferring the capital shareholding ratio of the bullish trend. Another object of the present invention is to provide an AI method for analyzing the long trend of the stock market. The AI method can be used to adaptively analyze the capital ratio of the current long trend of the stock market by reorganizing the fuzzy logic rule base from time to time as an artificial intelligence system. Basis of judgment and treatment. [Invention Description] (Please read the notes on the back before filling in this page) · Install ·

本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 A7 B7 五、發明說明(5 ) 有鑑於股票市場中資訊指標資料眾多,投資者無法一 一追蹤與監測資料訊息的變化,並且容易受到市場消息面 的影響而造成判斷上的錯誤,本發明遂提供一種分析股市 多頭走勢的多項技術指標之間的關係,以獲得投資資金與 作多持股資金比例之AI方法,作爲人工智慧判斷之依據。 本發明係爲一用於分析具有投資風險的資訊分析方 法,尤其是指股票市場的資訊,其主要的技術特徵係從股 票巾場的歷史資訊中找出過去發生的多頭走勢區間,這些 多頭走勢區間的界定是在一大盤底部區間確定後到下一 波的大盤頭部區間的走勢,藉以根據在多頭走勢區間的初 期應掌握較多的作多資金以獲得較大的獲利,而多頭走勢 區間的後期應降低作多資金以作爲風險控管,本發明利用 在多頭走勢區間的初期與後期之時段組織對應時段內的 技術指標間關係(Index-relation),例如:K値>〇値、 +DIF>-DIF等,並且以不同的技術指標間關係之組合作爲 一模糊邏輯規則的條件組合,從而在股票市場的歷史資訊 建立以模糊邏輯分析技術指標間關係的模糊邏輯規則 庫,俾使由模糊邏輯規則庫來推論在大盤底部區間確定 後,目前投資者應掌握的投資資金與作多資金的比例。 在本發明的較佳實施例中,以過去股票市場中大盤指 數的歷史資料,根據確立的大盤頭部區間與大盤底部區 間,以找出股市多頭走勢區間,該多頭走勢區間經一預設 的時間分割比例以界定出比例事件的時段(Time slot),根 據各比例事件對應的時段,從多項技術指標,例如:隨機 (請先閱讀背面之注意事項再填寫本頁) --------^---------痒, 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS)A4規格(21〇 x 297公釐) 573267 A7 B7 五、發明說明(6) 指標(KD値)、平滑異同平均線(MACD)、相對強弱指標 (RSI)、心理線(PSY)、動量指標(MTM)、乖離率(BIA)、騰 落指標(ADL)、漲跌比率(ADR)、逆勢操作(CDP)、振盪指 標(OSC)、…等之中界定出彼此的關係,再從該技術指標 間關係(index-relation)的歷史資料中自動組織各比例事件 的條件集合,並從各比例事件的條件集合中建立各技術指 標間關係與資金持股比例的歸屬函數,從而建立模糊邏輯 規則庫,俾使目前大盤指數在確立大盤頭部區間進入多頭 走勢後,根據前述模糊邏輯規則庫推論出現階段資金持股 比例。 本發明股市多頭走勢資金持股比例之AI方法及其諸 多優點與特徵將從下述詳細說明及所附圖式中得到進一 步的瞭解。 [圖式標號說明] (請先閱讀背面之注意事項再填寫本頁) 訂· 經濟部智慧財產局員工消費合作社印製 成交量 1 大盤指數 2 季線MAP72d 3 半年線map6M 4 年線map72M 5 成交量/日均線走勢mav,d 11 頭部量曲線 12 底部量曲線 13 大盤指數/日均線走勢map/d 21 [發明之詳細說明] 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公t ) 573267 A7 _B7 五、發明說明(7 ) 首先參考圖一所示,爲股票市場資訊之大盤指數與成 交量的走勢圖。股票市場資訊的大盤指數與成交量是每個 投資者所要掌握的重要訊息。在圖一的走勢圖中,大盤指 數與成交量彼此在時間軸上是一一對應,並且依時間軸的 分類有「日線走勢圖」、「週線走勢圖」與「月線走勢圖」。 在日線走勢圖中,每筆資料以「一日」爲基礎,包含當曰 股市大盤開盤指數、大盤收盤指數、當日大盤最高指數與 當日大盤最低指數,並對應到當日成交量,而週線走勢圖 與月線走勢圖則分別以「一週」與「一月」爲基礎的資料。 此外,走勢圖中常被投資者所觀察的尙包含均線,如6曰 均線、12日均線、24日均線與季線(72日均線)等,用以 判斷指數的趨勢。 投資者在觀察圖一所示的走勢圖中,最想掌握的訊息 莫過於大盤指數之「頭部區間」與「底部區間」,因爲這 對投資者的投資環境而言是處於「利空出盡」與「利多出 盡」的狀況。大盤頭部與底部的形成往往反應在成交量的 變化上,而對投資者是很難在這麼龐大的數據資料中找出 關聯性,即使,憑藉著個人所累積的經驗,也很難即時地 觀察預測出形成頭部與底部的徵兆。所以,本發明方法將 在大盤指數確立「底部區間」進入多頭走勢後,以建立的 模糊邏輯規則庫推論出現階段資金持股比例。 本發明方法,首先在過去股票市場中從大盤指數走勢 圖的歷史資料,界定出大盤「頭部區間」與「底部區間」, 藉以獲得大盤底部走向頭部的「多頭走勢」;根據界定的 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁)This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 573267 A7 B7 V. Invention Description (5) In view of the large number of information indicators in the stock market, investors cannot track and monitor the information Changes, and are susceptible to judgment errors caused by the influence of market news, the present invention then provides an AI method to analyze the relationship between multiple technical indicators of the stock market bull trend, in order to obtain investment funds and the proportion of long-held funds, As the basis for artificial intelligence judgment. The present invention is an information analysis method for analyzing investment risk, especially the information of the stock market. Its main technical feature is to find out the past long trend interval from the historical information of the stock market. These long trends The definition of the interval is the trend of the head of the broader market after the determination of the bottom interval of a large market, so that according to the initial period of the long trend interval, more long funds should be mastered to obtain greater profits, while the long trend In the later period of the interval, multiple funds should be reduced to be used as risk control. The present invention uses the index-relation in the corresponding period to organize the corresponding period in the early and late periods of the bull's trend interval, for example: K 値 > 〇 値, + DIF > -DIF, etc., and the combination of the relationship between different technical indicators as a conditional combination of fuzzy logic rules, so as to establish a fuzzy logic rule base based on fuzzy logic analysis of the relationship between technical indicators in the historical information of the stock market, 俾After the fuzzy logic rule base is used to infer that the interval at the bottom of the market is determined, the current investment funds that investors should have Proportion of long funds. In a preferred embodiment of the present invention, based on the historical data of the broad market index in the stock market in the past, according to the established broad market head interval and broad market bottom interval, the stock market bull trend interval is found, and the bull trend interval is preset. The time division ratio defines the time slot of the proportional event. According to the time period corresponding to each proportional event, from a number of technical indicators, such as: random (please read the precautions on the back before filling this page) ------ -^ --------- Itching, printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, this paper is printed in accordance with Chinese National Standard (CNS) A4 (21 × x297 mm) 573267 A7 B7 V. Invention Explanation (6) indicator (KD 値), smooth similarity average (MACD), relative strength indicator (RSI), psychological line (PSY), momentum indicator (MTM), deviation rate (BIA), take-off indicator (ADL), rising The relationship between each other is defined in the drop ratio (ADR), contrarian operation (CDP), oscillator index (OSC), etc., and then the proportional events are automatically organized from the historical data of the technical index-relation. Condition set, and the bar from each scale event The attribution function of the relationship between various technical indicators and the proportion of funds held in the set is established, so as to establish a fuzzy logic rule base, so that after the current broad market index establishes the head of the broad market and enters the bullish trend, it is inferred according to the aforementioned fuzzy logic rule base that funds appear in the stage Shareholding ratio. The AI method of the long-term capital holding ratio of the stock market of the present invention and its many advantages and characteristics will be further understood from the following detailed description and the attached drawings. [Illustration of Symbols] (Please read the notes on the back before filling out this page) Order · Printed by the Intellectual Property Bureau Staff Consumer Cooperative of the Ministry of Economic Affairs 1 Volume index 2 Quarterly line MAP72d 3 Half-year line map6M 4 Year-line map72M 5 Deals Volume / day moving average trend mav, d 11 Head volume curve 12 Bottom volume curve 13 Market index / day moving average trend map / d 21 [Detailed description of the invention] This paper scale applies Chinese National Standard (CNS) A4 specifications (210 X 297 (T) 573267 A7 _B7 V. Description of the Invention (7) First, referring to Figure 1, it is a trend chart of the broad market index and trading volume of stock market information. The broad market index and trading volume of stock market information are important information for every investor. In the chart in Figure 1, the market index and trading volume correspond to each other on the time axis, and according to the classification of the time axis, there are "daily chart", "weekly chart" and "monthly chart" . In the daily chart, each piece of data is based on "one day", including the opening index of the stock market, the closing index of the market, the highest index of the day, and the lowest index of the day, and corresponds to the trading volume of the day. The trend chart and the monthly trend chart are based on "week" and "January" data, respectively. In addition, the charts that are often observed by investors include moving averages, such as the 6-day moving average, the 12-day moving average, the 24-day moving average, and the quarterly line (72-day moving average). They are used to judge the trend of the index. When investors observe the trend chart shown in Figure 1, the most important information for investors to grasp is the "head range" and "bottom range" of the broad market index, because this is in a "bad sell out" for investors' investment environment. "And" Lido is running out. " The formation of the market's head and bottom often reflects the change in trading volume, and it is difficult for investors to find the correlation in such huge data. Even with the accumulated experience of individuals, it is difficult to instantly Observation predicted signs of head and bottom formation. Therefore, the method of the present invention will use the established fuzzy logic rule base to infer the proportion of funds held in the emergence stage after the "bottom range" of the broad market index is established and enters the bull trend. In the method of the present invention, first of all, from the historical data of the stock market index chart in the past stock market, the "head range" and "bottom range" of the broad market are defined, so as to obtain the "long trend" of the bottom of the broad market toward the head. Paper size applies to China National Standard (CNS) A4 (210 X 297 mm) (Please read the precautions on the back before filling this page)

Order

經濟部智慧財產局員工消費合作社印製 573267 A7 B7 五、發明說明(8 ) (請先閱讀背面之注意事項再填寫本頁) 大盤多頭走勢區間,找出對應該大盤多頭走勢區間內各項 技術指標的數據資料,例如:隨機指標(KD値)、平滑異同 平均線(MACD)、相對強弱指標(RSI)、心理線(PSY)、動量 指標(MTM)、乖離率(BIA)、騰落指標(ADL)、漲跌比率 (ADR)、逆勢操作(CDP)、振盪指標(OSC)、…等技術指標, 且該多頭走勢區間經一預設的時間分割比例以界定出複 數個比例事件的時段,而由每個比例事件組織各項技術指 標在對應時段內的資料,由各項技術指標間關係之組合作 爲每一比例事件的模糊邏輯規則的條件集合,據以建立起 各項技術指標間關係的模糊子集合(Fuzzy Subset),且其歸 屬函數則由對應時段內收集的資料來界定。 經濟部智慧財產局員工消費合作社印製 根據本發明方法,當建立比例事件的模糊邏輯規則 後,一集合合倂程序被應用於各項技術指標間關係的模糊 子集合,將一些歸屬函數分布接近的模糊子集合加以合 倂,以減少各項技術指標的模糊子集合數量,使模糊邏輯 規則庫有效地規範各比例事件,以提昇模糊邏輯規則推論 的速度與推論的效率。俾使在大盤指數確立底部區間走入 多頭市場中,由前述模糊邏輯規則庫推論現階段各項技術 指標間關係的條件,以決定目前投資者應採取的資金持股 比例。 在本發明的較佳實施例中,大盤指數日線走勢、週線 走勢與月線走勢的資料係由收盤指數所組成,分別由 PdGd)、Pw(iw)與PmGm)所表示,其中,iD以每日爲單位; iw以每週爲單位;心以每月爲單位。此外,成交量日線走 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 ⑴ A7 B7 五、發明說明(9 ) 勢、週線走勢與月線走勢的資料係由累積成交量所組成, 分別由VD(iD)、^~)與Vm(心)所表示,而各項技術指標 的資料如乖離率可分別由BIAD(iD)、BIAw(iw)與ΒΙΑΜ(ίΜ) 表示日線走勢、週線走勢與月線走勢的資料,或相對強弱 指標可分別由RSID(iD)、RSIw(iw)與RSIM(iM)所表示,依 此類推,其他技術指標的資料表示。 所以,大盤指數/日均線走勢MAP/D的資料由下式(1) 表示: Σ^ο-h) h=0 由上式⑴侍知’ MAP/dQd) = Pd(广d) 而成交量/日均線走勢MAViO的資料由下式(2)表示 (請先閱讀背面之注意事項再填寫本頁) ⑺ MAV/D⑹ Σ^-h) /?=0 經濟部智慧財產局員工消費合作社印製 由上式(2)得知,MAV川(iD) = VD(iD) 請參考圖二,爲本發明方法中界定大盤指數之頭部區 間的示意圖。本發明方法從股市資訊的歷史資料PD(/D)、 Pw(M、PM(〜)或MAP/D(iD)中,界定大盤指數之頭部區間 的程序包含: a)以ζ·日均線走勢MAP/d(〖d)爲基礎,在時間軸取一* 時間區間T, 其中,在時間區間T內,大盤指數之季線map72D或 半年線map6M或年線map/2M爲下凹曲線,即 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 A7 B7 五、發明說明(10) MAP72D={,d I Zmax=max MAP72D〇D), 不爲T的端點,iDGT} MAP福={ίΜ I Zmax=max ΜΑΡ_ (ίΜ), iM不爲T的端點,〖MeT} MAP/2M=<Um I Zmax^inax MAP/2M(广Μ), 以不爲τ的端點,iMeT} (3) 而且,在時間區間Τ內,/日均線走勢MAP/D(iD)具有 至少一個極大値Zm與至少一個極小値zn,而最大値Zmax 爲極大値冗,。之一,即 MAP/D={iD,m,n| Zm=local—maxMAP/D(iD)且Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 573267 A7 B7 V. Invention Description (8) (Please read the notes on the back before filling this page) The bull market trend range, find out the various technologies in the bull market trend range Indicator data, such as: Stochastic indicator (KD 値), Smoothed Similarity and Average (MACD), Relative Strength Index (RSI), Psychological Line (PSY), Momentum Index (MTM), Deviation Rate (BIA), Rising Index ( (ADL), up-down ratio (ADR), contrarian operation (CDP), oscillation indicator (OSC), ... and other technical indicators, and the long trend interval is divided by a preset time to define the period of multiple proportional events The data of each technical indicator in the corresponding period is organized by each proportional event, and the combination of the relationships between the technical indicators is used as the condition set of the fuzzy logic rules of each proportional event, so as to establish the various technical indicators. The fuzzy subset of the relationship, and its attribution function is defined by the data collected during the corresponding period. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. According to the method of the present invention, after the fuzzy logic rule of proportional events is established, a set of combining procedures is applied to the fuzzy sub-set of the relationship between various technical indicators, and some attribution functions are distributed close to The fuzzy sub-sets are combined to reduce the number of fuzzy sub-sets of various technical indicators, so that the fuzzy logic rule base can effectively regulate the proportional events, so as to improve the speed and efficiency of inference of fuzzy logic rules. In order to enter the long market when the bottom index of the broad market index is established, the conditions of the relationship between various technical indicators at this stage are deduced from the aforementioned fuzzy logic rule base to determine the proportion of funds that investors should take. In a preferred embodiment of the present invention, the data of daily, weekly and monthly trends of the broad market index are composed of closing indices, which are represented by PdGd), Pw (iw), and PmGm, respectively, where iD In daily units; iw in weekly units; heart in monthly units. In addition, the daily trading volume of the paper applies the Chinese national standard (CNS) A4 specification (210 X 297 mm) 573267 ⑴ A7 B7 V. Description of the invention (9) Trend, weekly trend and monthly trend data are provided by The accumulated trading volume is represented by VD (iD), ^ ~), and Vm (heart), and the data of various technical indicators such as deviation rate can be respectively determined by BIAD (iD), BIAw (iw), and ΒΙΑΜ (ίΜ ) Data representing daily, weekly and monthly trends, or relative strength indicators can be represented by RSID (iD), RSIw (iw), and RSIM (iM) respectively, and so on, and other technical indicators' data . Therefore, the data of the market index / day moving average trend MAP / D is expressed by the following formula (1): Σ ^ ο-h) h = 0 by the above formula ⑴ 知 知 'MAP / dQd) = Pd (广 d) and the trading volume / The data of the daily moving average MAViO is expressed by the following formula (2) (Please read the precautions on the back before filling in this page) ⑺ MAV / D⑹ Σ ^ -h) /? = 0 Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs It is known from the above formula (2) that MAV Chuan (iD) = VD (iD) Please refer to FIG. 2, which is a schematic diagram for defining a head interval of a broad market index in the method of the present invention. In the method of the present invention, from the historical data PD (/ D), Pw (M, PM (~), or MAP / D (iD) of the stock market information, the procedure for defining the head range of the broad market index includes: a) the ζ · day moving average Based on the trend MAP / d (〖d), take a * time interval T on the time axis, where within the time interval T, the seasonal index map72D of the broad market index or the semi-annual map6M or annual map / 2M is a concave curve, That is, this paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 573267 A7 B7 V. Description of the invention (10) MAP72D = {, d I Zmax = max MAP72D〇D), not the endpoint of T , IDGT} MAP Fu = {ίΜ I Zmax = max ΜΑΡ_ (ίΜ), iM is not the endpoint of T, 〖MeT} MAP / 2M = < Um I Zmax ^ inax MAP / 2M (广 Μ), so End point of τ, iMeT} (3) Moreover, in the time interval T, the / day moving average trend MAP / D (iD) has at least one maximum 値 Zm and at least one minimum 値 zn, and the maximum 値 Zmax is a maximum redundancy, . One, that is, MAP / D = {iD, m, n | Zm = local—maxMAP / D (iD) and

Zn二local—min MAP/dGd),beT} (4) b)決定一α値,以獲得一連續之時間區間Ττ,使得 MAP,d 存在 MAP/D(iD)2a,iDeTT 且 MAP/D(iD)<a,iD芒Ττ, 其中,α値選自程序a)中極小値ζη之一,即 MAP/D(iD)<a,ίνΤτ 且 aezn } (5) 則時間區間Ττ爲大盤指數之頭部區間。 根據本發明的較佳實施例,在界定大盤指數之頭部區 間的程序中,程序a)的時間區間Τ可選自半年到一年之 間,或選自7個月到12個月之間,或最好選自30週到46 週之間,而以/日均線走勢MAP/D(iD)爲基礎,通常MAP/D(iD) 係以MAP3D、MAP6D、MAP—或 MAPwD爲佳;而程序b) 中獲得的連續之時間區間Ττ一般是介於7日到21日之間 爲佳,或最好是大約兩週左右,藉以決定出大盤指數之頭 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製Zn 2 local-min MAP / dGd), beT} (4) b) Determine an α 値 to obtain a continuous time interval Tτ, so that MAP, d exists in MAP / D (iD) 2a, iDeTT and MAP / D ( iD) < a, iD ττ, where α 値 is selected from one of the minimum 値 ζη in the program a), that is, MAP / D (iD) < a, ίνττ and aezn} (5) then the time interval ττ is the broader market Index head range. According to a preferred embodiment of the present invention, in the procedure for defining the head interval of the broad market index, the time interval T of procedure a) may be selected from half a year to one year, or from 7 months to 12 months , Or preferably selected between 30 weeks and 46 weeks, and based on the MAP / D (iD) of the daily average movement, usually MAP / D (iD) is preferably MAP3D, MAP6D, MAP—or MAPwD; and The continuous time interval Tτ obtained in Procedure b) is generally between 7th and 21st, or preferably about two weeks, so as to determine the head of the broader market index. The paper size applies the Chinese national standard (CNS ) A4 size (210 X 297 mm) (Please read the notes on the back before filling this page) Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs

573267 A7 B7 五、發明說明(11) 部區間Ττ。 接著請參考圖三’爲本發明方法中界定大盤指數之底 部區間的示意圖。本發明方法從股市資訊的歷史資料 PdGd)、Pw(〖w)、Ρμ(?μ)或MAP/D(iD)中’界定大盤指數之 底部區間的程序包含: a) 以ζ·日均線走勢MAP/D(iD)爲基礎,在時間軸取一 時間區間T, 其中,在時間區間T內,大盤指數之季線MAP72D或 半年線map6M或年線map/2M爲上凹曲線,即 MAP72D={,D I zmin=min MAP72D〇D), 化不爲T的端點,iDeT} MAP— ={,M I Zmin=min MAP— (ίΜ), 心不爲T的ii而點,心^丁} MAP;2M~{^M I Zmin=lTlin MAP72m(^m)? iM不爲T的端點,(6) 而且,在時間區間τ內,/日均線走勢MAPiD(iD)具有 至少一個極大値Zm與至少一個極小値zn,而最小値Zmin 爲極小値zn之一,即 MAPzD={/D,m,n| Zm=local—maxMAP/D(iD)且573267 A7 B7 V. Description of the invention (11) Part interval Tτ. Please refer to FIG. 3 'for a schematic diagram of defining a bottom range of a broad market index in the method of the present invention. In the method of the present invention, the procedure of defining the bottom range of the broad market index from the historical data of stock market information (PdGd), Pw (〖w), Pμ (? Μ) or MAP / D (iD) includes: a) trending with ζ · day moving average Based on MAP / D (iD), a time interval T is taken on the time axis. In the time interval T, the seasonal index MAP72D or semi-annual line map6M or annual line map / 2M of the broad market index is a concave curve, that is, MAP72D = {, DI zmin = min MAP72D0D), the end point that is not T, iDeT} MAP — = {, MI Zmin = min MAP— (ίΜ), the heart is not the point of T, and the heart is} MAP ; 2M ~ {^ MI Zmin = lTlin MAP72m (^ m)? IM is not the endpoint of T, (6) Moreover, in the time interval τ, the average daily trend MAPiD (iD) has at least one maximum 値 Zm and at least A minimum 値 zn, and the minimum 値 Zmin is one of the minimum 値 zn, that is, MAPzD = {/ D, m, n | Zm = local—maxMAP / D (iD) and

Zn=local一min MAP/D(iD),ί#Τ} (7) b) 決定一β値,以獲得一連續之時間區間ΤΒ,使得 MAP/D 存在 MAP/D(iD)邻,iDeTB 且 ΜΑΡ/Ό(ί〇)<β,ίΜΤΒ, 其中,β値選自程序a)中極大値zm之一,即 MAP/D= {iD,m| 邛,TT=>MAP/D(iD)邻,iDeTB 且 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 裝 訂· 經濟部智慧財產局員工消費合作社印製 573267 A7 B7 i、發明說明(12) ΜΑΡω(ίΐ))<β,iD 茫 Tb 且 } ⑻ 則時間區間TB爲大盤指數之底部區間。 根據本發明的較佳實施例,在界定大盤指數之底部區 間的程序中,程序a)的時間區間T可選自半年到一年之 間,或選自7個月到12個月之間,或最好選自30週到46 週之間,而以/日均線走勢MAPzD(iD)爲基礎,通常MAP/D(iD) 係以MAPjd、MAP5D、MAP72D或MAP㈣爲佳;而程序b) 中獲得的連續之時間區間Ττ一般是介於7日到21日之間 爲佳,或最好是大約兩週左右,藉以決定出大盤指數之底 部區間Τβ。 · 根據本發明方法上述的程序,可在歷史資料的時間軸 上決定出複數個頭部區間Ττι、τΤ2…與複數個底部區間 TB1、ΤΒ2…,因此可以進一步決定出複數個在確立大盤底 部區間τΒ的多頭走勢區間TBullish。該多頭走勢區間TBullish 係在底部區間τΒ至下一個相鄰之頭部區間Ττ期間,而由 確立大盤底部區間ΤΒ之時點到該頭部區間Ττ中最大値 zmax之時點所界定,βρ TBullish^t ?Dmax ]5 . (9) 其中 MAP/D⑺冰 for iDmax,且 MAp▲瞧)=^ 因此,由上式(9)可知,多頭走勢區間TBuiiish的成立係 大盤均線走勢ΜΑΡω⑺突破在大盤底部區間的頸線β之 後。對於確立股市進入多頭走勢區間TBuiUsh之初,投資者 應有效的將掌握作多資金投注在作多股票上,而隨著多頭 走勢區間TBuliish變化走入尾段期間或大般均線走勢 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公f ) (請先閱讀背面之注意事項再填寫本頁) ta.Zn = local-min MAP / D (iD), ί # Τ} (7) b) Determine a β 値 to obtain a continuous time interval TB, so that MAP / D has MAP / D (iD) neighbors, iDeTB and ΜΑΡ / Ό (ί〇) < β, ίΜΤΒ, where β 値 is selected from one of the largest 値 zm in the program a), that is, MAP / D = {iD, m | 邛, TT = > MAP / D (iD ) Neighbor, iDeTB, and this paper size applies Chinese National Standard (CNS) A4 specification (210 X 297 mm) (Please read the precautions on the back before filling this page) Binding · Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs and Consumer Cooperatives 573267 A7 B7 i. Explanation of the invention (12) ΜΑΡω (ίΐ)) < β, iD TTb and} 则 Then the time interval TB is the bottom interval of the broad market index. According to a preferred embodiment of the present invention, in the procedure for defining the bottom interval of the broad market index, the time interval T of procedure a) may be selected from half a year to one year, or from 7 months to 12 months, Or preferably from 30 to 46 weeks, and based on the MAPzD (iD) of the daily average movement, usually MAP / D (iD) is preferably MAPjd, MAP5D, MAP72D, or MAP㈣; and program b) The obtained continuous time interval Tτ is generally between 7th and 21st, or preferably about two weeks, so as to determine the bottom interval Tβ of the broad market index. · According to the above-mentioned program of the method of the present invention, a plurality of head intervals ττι, τΤ2, ... and a plurality of bottom intervals TB1, TB2, ... can be determined on the time axis of historical data, so it can be further determined that a plurality of bottom intervals are established The long range of τΒ is TBullish. The long trend interval TBullish is between the bottom interval τB and the next adjacent head interval ττ, and is defined by the time point at which the bottom interval TB of the broader market is established to the time point of the maximum 値 zmax in the head interval ττ, βρ TBullish ^ t ? Dmax] 5. (9) where MAP / D⑺ice for iDmax, and MAp ▲) = ^ Therefore, from the above formula (9), we can see that the establishment of the long trend interval TBuiiish is the market moving average trend ΜΑΡω⑺ breaks through the bottom of the market. After the neckline β. At the beginning of the establishment of the stock market entering the long trend range TBuiUsh, investors should effectively bet on long funds to bet on long stocks, and as the long trend range TBuliish changes into the tail period or the general average trend, this paper applies China National Standard (CNS) A4 Specification (210 X 297 male f) (Please read the precautions on the back before filling this page) ta.

經濟部智慧財產局員工消費合作社印製 573267 Α7 ______ Β7 經濟部智慧財產局員工消費合作社印製 五、發明說明(13) MAP,D⑺可能走入下一個相鄰之頭部區間ττ期間,投資者 應要能逐步因應盤中技術指標間關係有效調節資金,將部 份作多持股「獲利了結」,避免在多空反轉之際,手上有 過多的持股,使作多資金被套牢。 根據本發明方法,爲了在多頭走勢區間TBullish適應盤 中技術指標間關係,在各項技術指標中界定出多頭走勢區 間TBumsh各項技術指標間關係,再從股市歷史資料以自動 方式組織技術指標間關係的模糊邏輯規則庫,因此在多頭 走勢區間TBullish分割出資金持股比例事件的時段,藉以根 據比例事件的各個時段組織出模糊邏輯規則。請一起參考 圖四,顯示了用以分割資金持股比例事件的時段的時間分 割比例圖。 在本發明的一種實施例中,如圖四所示,在一多頭走 勢區間TBullish之內,投資者的資金持股比例應在區間 TBullish初段將較多的資金轉換投資在作多的股票上,所以 通常剩餘資金與持股資金的比例,例如爲1 : 9,並且隨著 技術指標間關係的變化,大盤均線走勢MAP/D⑺可能進入 不同時段,而投資者在風險的考量下應適度調整資金持股 比例;倘若技術指標間關係的變化反應指出,多頭走勢區 間TBullish可能走入尾段或大盤均線走勢MAP/D⑺可能走入 下一個相鄰之頭部區間Ττ時,投資者應逐步調節剩餘資金 與持股資金的比例,例如爲9 : 1 ’將作多的股票「獲利了 結」。 所以,根據本發明的實施例,如圖四所示的時間分割 (請先閱讀背面之注意事項再填寫本頁) 裝 訂·Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 573267 Α7 ______ Β7 Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs V. Invention Description (13) MAP, D⑺ may enter the next adjacent head interval period ττ. It should be able to gradually adjust the funds in response to the relationship between the technical indicators in the intraday market, and "make a profit" on part of the long holdings to avoid excessive holdings in the hands of the long and short reversals, which will make the long funds Hold up. According to the method of the present invention, in order to adapt the relationship between intraday technical indicators in the long trend interval TBullish, the relationship between the technical indicators of the long trend interval TBumsh is defined in each technical indicator, and then the technical indicators are automatically organized from the historical data of the stock market. The fuzzy logic rule base of the relationship, so in the long trend interval TBullish, the period of the fund holding proportion event is divided, so as to organize the fuzzy logic rule according to each period of the proportional event. Please refer to Figure 4 together, which shows the time division ratio chart for the period used to divide the capital shareholding event. In an embodiment of the present invention, as shown in FIG. 4, within a long trend interval TBullish, the investor's capital shareholding ratio should be converted to invest in more stocks at the beginning of the interval TBullish Therefore, the ratio of the remaining funds to the shareholding funds is usually 1: 9, and as the relationship between technical indicators changes, the broader market moving average trend MAP / D⑺ may enter different periods, and investors should adjust appropriately under the consideration of risk Shareholding ratio of funds; if the change in the relationship between technical indicators indicates that the bullish trend interval TBullish may enter the end or the broader moving average trend MAP / D⑺ may enter the next adjacent head interval ττ, investors should gradually adjust The ratio of the remaining funds to the shareholding funds is, for example, 9: 1 'the long stock will be "profitable". Therefore, according to the embodiment of the present invention, the time division is shown in FIG. 4 (please read the precautions on the back before filling this page)

本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明說明(14) 比例圖,投資者在多頭走勢區間TBullish期間大盤均線走勢 MAP/D⑺走入下一個相鄰之頭部區間Ττ過程中,剩餘資金 與作多資金的比例,可能適應各時段例技術指標資訊的變 化,由1 : 9的比例逐步調整到9 : 1。換言之,投資者在 多頭走勢區間TBullish期間逐步將作多的股票「獲利了結」, 使投資風險降低。 根據本發明的不同實施例,時間分割比例可界定出不 同的時段長度與資金持股比例値,而一時段長度對應一資 金持股比例値爲一比例事件E。繼續參考圖四,在本發明 此一實施例中,時間分割比例在一多頭走勢區間丁^㈣界 定出九個比例事件E“=l to 9,每一比例事件E/的資金持 股比例値對應其時段^而從區間初段到尾段由1: 9逐步調 整到9 : 1。因此,從股市歷史資料中界定出的每一多頭走 勢區間TBullish便可根據時間分割比例的時段,組織每—比 例事件艮的模糊邏輯規則,分別如下列所示:This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 573267 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 B7 V. Description of the invention (14) Proportion map, investors during the long trend interval TBullish period During the course of the market moving average MAP / D⑺ entering the next adjacent head interval Tτ, the ratio of remaining funds to long funds may be adapted to the changes in technical indicators and information in each period, from a 1: 9 ratio to 9 : 1. In other words, during the bullish trend interval TBullish, investors gradually "profitable" long stocks, reducing investment risk. According to different embodiments of the present invention, the time division ratio can define different lengths of time periods and the capital shareholding ratio 値, and a period length corresponding to a capital shareholding ratio 値 is a proportional event E. Continuing to refer to FIG. 4, in this embodiment of the present invention, the time division ratio is defined in a long trend interval. Nine ratio events E “= l to 9 are defined, and each ratio event E / is the capital shareholding ratio.値 Corresponding to its time period ^ and gradually adjusted from 1: 9 to 9: 1 from the beginning to the end of the interval. Therefore, each bull trend interval TBullish defined from the stock market historical data can be divided into proportions of time according to time, organization The fuzzy logic rules of each-proportional event are as follows:

Rule—E/ : IF indexK>D is 7^>D and zWex+DI>.DI [s ΛRule-E /: IF indexK > D is 7 ^ > D and zWex + DI > .DI [s Λ

+DI>-DI aiIU indexO\¥>〇 is ^IF>〇5 for tetj. THEN Ratio is 9:1.+ DI > -DI aiIU indexO \ ¥ > 〇 is ^ IF > 〇5 for tetj. THEN Ratio is 9: 1.

RuleJE2 :RuleJE2:

IF indexK>D is and index^D]>_Dl is F2 anH + D1>-DI uu m^xDiF>〇 is ^1ρ>〇? for THEN Ratio is 8:2.IF indexK > D is and index ^ D] > _Dl is F2 anH + D1 > -DI uu m ^ xDiF > 〇 is ^ 1ρ > 〇? For THEN Ratio is 8: 2.

Rule_E3: 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) --------------裝--- (請先閱讀背面之注意事項再填寫本頁) · --線 573267 A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明說明(15) IF indexK>O is 」 iyidex^Y^Q is THEN Ratio is 7:3.Rule_E3: This paper size is applicable to China National Standard (CNS) A4 (210 X 297 mm) -------------- install --- (Please read the precautions on the back before filling in this (Page) ·-573267 A7 B7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs V. Invention Description (15) IF indexK > O is '' iyidex ^ Y ^ Q is THEN Ratio is 7: 3.

73 K>D and index+^\>^] is jP ;;F>〇? for tet3.73 K > D and index + ^ \ > ^] is jP ;; F > 〇? For tet3.

3 + DI>-DI and3 + DI > -DI and

Rule一E〆 IF indexK>r> is FK4>D and index^m>.m is F indexD]F>〇 is Fd4]F>〇5 for tet4. THEN Ratio is 6:4. RuleJE5 : IF index^>D is FK5>D and index+D]>.O] is F5〇I>DI indexDiF>〇 is Fd5if>〇? for THEN is 5:5. Rule_E6 ·· IF indexK>D is FK6>D and z/7^x+DI>.DI is F6DI>DI and indexDip>〇 is ^DIF>0? for tet^· THEN is 4:6. Rule_E7 · IF index\<i>O is F:O and z^Jex+Di>-Di is F:O,〜and indexDiF>〇 is FJif>〇5 for tet7. THEN is 3:7.Rule-E〆IF indexK > r > is FK4 > D and index ^ m > .m is F indexD] F > 〇is Fd4] F > 〇5 for tet4. THEN Ratio is 6: 4. RuleJE5: IF index ^ > D is FK5 > D and index + D] > .O] is F5〇I > DI indexDiF > 〇is Fd5if > 〇? For THEN is 5: 5. Rule_E6 ·· IF indexK > D is FK6 > D and z /7^x+DI>.DI is F6DI > DI and indexDip > 〇is ^ DIF > 0? For tet ^ · THEN is 4: 6. Rule_E7 · IF index \ < i > O is F: O and z ^ Jex + Di > -Di is F: O, ~ and indexDiF > 〇is FJif > 〇5 for tet7. THEN is 3: 7.

ule Eule E

4 + DI>-DI and and (請先閱讀背面之注意事項再填寫本頁) · 丨線4 + DI > -DI and and (Please read the notes on the back before filling this page)

+DI>-DI ck>d is Fk8>d and index^m>.m is F8 zWexDiF>0 is i7二〇, for IEN Ratio is 2:8. indexy index+ DI > -DI ck > d is Fk8 > d and index ^ m > .m is F8 zWexDiF > 0 is i7 two zero, for IEN Ratio is 2: 8. Indexy index

k>d !Sk > d! S

!?9 K>D!? 9 K > D

+ DI>-DI and and index^D]>.Di is F9〇] 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 A7 i、發明說明(16) THEN Ratio is 1:9. 其中,在上述模糊邏輯規則中,M办Xk>D、办X+DI>-DI 與/m/exDIF>Q分別表示邏輯規則的技術指標間關係條件集 合的語意變數,而i^>D、係對應各語意變數 經濟部智慧財產局員工消費合作社印製 Z>^eXK>D、k(^X+Di>_DI、/mfecDIF>0 的模糊子集合;此夕t,如加 則爲邏輯規則的資金持股比例推論結果的語意變數。需進 一步地說明,在此一實施例中所組織的模糊邏輯規則僅爲 一範例,邏輯規則的技術指標間關係條件集合除了 K>D, +DI>-DI,DIF>0之外,亦可爲K>80,…等技術指標間關係 的其一或多個的條件集合。 在每次從一多頭走勢區間TBullish中組織每一比例事件 E/的模糊邏輯規則,語意變數//7(^XK>D、—^x+DI>.DI、 /w/exDIF>()的模糊子集合便從多頭走勢區間TBullish的時間軸 上各時段所對應的技術指標資訊K(〇, D(〇, +DI⑺, -DI⑺,DIF⑺反應出各項技術指標間關係K>D、+DI>-DI、 DIF>0的程度,而觀察時間〖可以爲日線走勢、週線走勢 與月線走勢,以界定語意變數的歸屬函數,且該歸屬函數 μ的圖形係爲一三角形,如圖五所示,並由三個參數L, Μ 及決定之,如以下表示式: t — L Μ - L' 卜R R-M'+ DI > -DI and and index ^ D] &. Di is F9〇] This paper size is applicable to China National Standard (CNS) A4 (210 X 297 mm) 573267 A7 i. Description of the invention (16) THEN Ratio is 1: 9. Among the above fuzzy logic rules, M office Xk> D, office X + DI> -DI and / m / exDIF> Q respectively represent the semantic variables of the set of relationship conditions between the technical indicators of the logic rules, and i ^ > D is a fuzzy sub-set of Z > ^ eXK &D; d, k (^ X + Di > _DI, / mfecDIF > 0, corresponding to the semantic variables of the Intellectual Property Bureau of the Ministry of Economic Affairs; Plus, it is a semantic variable of the inference result of the logic of the capital shareholding ratio. It needs to be further explained that the fuzzy logic rule organized in this embodiment is only an example, and the set of relationship conditions between the technical indicators of the logic rule is in addition to K > In addition to D, + DI > -DI, DIF > 0, it may also be one or more condition sets of the relationship between technical indicators such as K > 80, .... Each time from the long trend interval TBullish is organized Fuzzy logic rule of a proportional event E /, semantic variable // 7 (^ XK > D, — ^ x + DI > .DI, / w / exDIF > () 's fuzzy sub-collection then reflects the technical indicator information K (〇, D (〇, + DI⑺, -DI⑺, DIF⑺) corresponding to each time period on the time axis of the long trend interval TBullish. K > D, + DI > -DI, DIF > 0, and the observation time may be daily trend, weekly trend, and monthly trend to define the attribution function of semantic variables, and the graph of the attribution function μ is It is a triangle, as shown in Figure 5, and is determined by three parameters L, Μ and, as shown below: t — L Μ-L 'BU R RM'

L<t<M M<t<R (10) 例如,在本發明的一種實施例中,從大盤6日均線走 勢MAP6D⑺的歷史資料獲得一多頭走勢區間TBullish,根據 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 訂·L < t < M M < t < R (10) For example, in one embodiment of the present invention, a long trend interval TBullish is obtained from historical data of the market's 6-day moving average trend MAP6D⑺, and the Chinese national standard is applied according to the paper scale (CNS) A4 specification (210 X 297 mm) (Please read the precautions on the back before filling this page) Order ·

573267 A7 B7 五、發明說明(π) 圖四所示的時間分割比例界定該多頭走勢區間TBullish的九 個比例事件E,,/=1 to 9,其中,模糊邏輯規則Rule_E/其 各語意變數、k^^+DI>-DI、k^^D_IF>G的模糊子集 合€>D,, 的資料集合分別爲 /^>d = {IQd⑺-Djd⑺},/Gh dDi={+Dl6D(0 -DldD⑺}, C>〇-{DIF^(〇}, tet] 因此,各模糊子集合圮>D,dDI,eF>。的歸屬函數界定 如下: μ^0(Ι?Μ^); I=min{ K5d(〇-D6d(〇 ) ? M=avg{ K6D(〇-I>6d(0} ? R=max{K(5D(〇-D(5D(i)}? teti μ:οι>—D,(L,關; Z=min{+DI6D(〇--DI6D(i)},M=avg{ +DI6d⑺一DI6d⑺), R=max{+Dl6D(i)—Dl6D(i)}? teti μ〇ΐΡ>〇(ΑΜ9τ?); L=minDIF5D(〇5 M=avgDIF^D(〇5 R=maxDIF(5D(〇5 tetj 其中,函數min係取資料集合中的最小値;函數 avg係取資料集合中的算術平均値;函數max係取iei/ 資料集合中的最大値。 依此類推,其他模糊邏輯規則Rule_E,的各語意變數 zWexK>D、"ito+Dl>_Di、zWexDIF>0戶斤對應的模糊子集合之歸 屬函數亦可依上述方式加以界定。因此,從大盤6日均線 走勢MAP6D(〇的歷史資料中所獲得的每一多頭走勢區間 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 訂·. 經濟部智慧財產局員工消費合作社印製 573267 A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明說明(18) TBullish可建立一完整的模糊邏輯規則庫,且輸入語意變數 可有系統地界定其模糊子集合對應技術指標資訊之歸屬 函數。此外,邏輯規則庫的資金持股比例推論結果的語意 變數/?_·〇之各比例的歸屬函數界定如下: L=4,M=9 μ:(Ι,Μ,7?); L=|,M=4,R=9 μ;1(Ζ,Μ,7?); L=!,M=!,R=4 Mw:(兑,乾穴);1,M=吾,R= I ^ lim 5 M= 1 ? R= j μ:1(Ι?Μ?/?); L=|?M=|?R=1ML)(厶,乾穴);>士,M=吾,R=. · μ;1(Ι5Μ57?); L=i?M=i?R=| C (-,μ,穴);Μ=去,R= 士 而圖六所示爲推論結果的語意變數之各比例的 歸屬函數圖。 由於從歷史資料中不同的多頭走勢區間TBullish所建立 的模糊邏輯規則庫,各輸入語意變數所組織的模糊子集合 數量眾多,所以本發明爲了提昇模糊邏輯規則庫推論的效 率,利用一合倂運算將重疊的模糊子集合加以合倂,以減 少輸入語意變數分割(partition)的模糊子集合數量。請配合 參考圖七,顯示重疊模糊子集合合倂之歸屬函數圖,所謂 的「重疊的模糊子集合」係指該等模糊子集合的歸屬函數 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 裝 . i線· 經濟部智慧財產局員工消費合作社印製 573267 A7 B7 五、發明說明(19) 在幾何位置上彼此有一定程度的相互重疊’而合倂運算係 將兩歸屬函數合倂成爲一個新的歸屬函數’以代表一新的 模糊子集合,取代原來兩模糊子集合在模糊邏輯規則的條 件組合,其合倂運算如下: 其中,Li=min(Uy),M丨,7?】=max((尺,矣) 至於兩歸屬函數合倂的條件,係兩歸屬函數的參數^ 落在一設定的範圍之內。在本發明的一種實施例中’歸屬 函數合倂的設定範圍係根據各表示的技術指標間關係而 定,且隨著該設定範圍動態地調大範圍’該語意'變數分割 的模糊子集合數量將有效地遞減。 因此,將各輸入語意變數的模糊子集合經過合倂運算 後,模糊邏輯規則庫的每一模糊邏輯規則將被重新檢視其 輸入語意變數的條件組合;在本發明的一種實施例中’倘 若有兩個或兩個以上的模糊邏輯規則在推論相同的資金 持股比例結果時,其輸入語意變數所對應的模糊子集合的 條件組合係相同,則該等模糊邏輯規則僅需保留一個模糊 邏輯規則於規則庫中。 在本發明的進一步實施例中,倘若有兩個或兩個以上 的模糊邏輯規則在推論相同的資金持股比例結果時,其輸 入語意變數所對應的模糊子集合的條件組合係不相同,則 該等模糊邏輯規則的條件組合彼此係邏輯「或」(OR)的關 係來推論相同的資金持股比例結果。 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 裝 線 573267 A7 B7 五、發明說明(20) 在本發明的再進一步實施例中,倘若有兩個或兩個以 上的模糊邏輯規則,其輸入語意變數所對應的模糊子集合 以相同條件組合推論不相同的資金持股比例結果時,則輸 出語意變數的歸屬函數執行合併運算,以推論出一 新的資金持股比例結果。 因此,根據本發明方法所建立的模糊邏輯規則庫,利 用模糊邏輯分析將各語意變數的技術指標間關係模糊化 處理(Fuzzification Process),經各模糊邏輯規則的推論 (Inference),將推論結果以解模糊化處理(Defuzzification Process)以獲得一資金持股比例或資金持股比例値;其 中,輸入的技術指標資訊經由其模糊子集合的歸屬函數以 模糊化處理,以及將推論的資金持股比例以例如 COA(Center of Area)、COS(Center of Sums)與 M0M(Mean of Maxima)的解模糊化處理,則已見於習知技藝的模糊邏 輯分析中。 最後請參考圖八,顯示本發明資金持股比例方法之流 程圖。本發明的方法係利用模糊邏輯分析股市多頭走勢以 推論投資者手上資金與投資在作多持股資金的比例關 係,該方法包含以下的程序: 在股市大盤走勢的歷史資料MAP/D(iD)中界定出大盤 頭部區間Ττ之集合,如圖二所示,而頭部區間Ττ的大盤 指數曲線由式(5)表示之; 在股市大盤走勢的歷史資料MAP/D(iD)中界定出大盤 底部區間TB之集合,如圖三所示,而底部區間TB的大盤 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 訂· --線· 經濟部智慧財產局員工消費合作社印製 573267 A7 B7 五、發明說明(21) 指數曲線由式(8)表不之; (請先閱讀背面之注意事項再填寫本頁) 從大盤頭部區間Ττ與大盤底部區間TB之集合界定出 股市大盤走勢的歷史資料中已確立大盤底部區間TB的多 頭走勢區間 ^Bullish 並由式(9)表示之; 根據一預設的時間分割比例,在每一多頭走勢區間 TBumsh界定出資金持股比例事件的時段,藉以組織資金持 股比例事件的模糊邏輯規則的條件組合,該條件組合係指 由各項技術指標資訊所反應的技術指標間關係之組合,而 由技術指標間關係之語意變數在指定時段中所對應的模 糊子集合,該模糊邏輯規則的推論結果則爲資金持股比例 値; 其中,由歷史資料的多頭走勢區間TBullish中所組織資 金持股比例事件的模糊邏輯規則庫,其模糊子集合的歸屬 函數係由該語意變數對應的技術指標間關係在對應的時 段中的資料所界定; 經濟部智慧財產局員工消費合作社印製 藉由一合倂運算將模糊邏輯規則庫各語意變數的模 糊子集合中,重疊的模糊子集合加以合倂,以減少該語意 變數分割的模糊子集合數量,而合倂運算由式(11)表示之; 進一步檢視模糊邏輯規則庫中,各模糊邏輯規則的條 件組合與推論結果,以移除規則庫用相同條件組合推論出 同一資金持股比例結果的模糊邏輯規則,減少該模糊邏輯 規則庫的多餘或重複的規則; 俾使本發明的方法以自動方式組織出模糊邏輯規則 庫,並降低規則庫的複雜度,以提供投資者若目前大盤走 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公f " 573267 A7 ____ B7 五、發明說明(22) (請先閱讀背面之注意事項再填寫本頁) 勢已確立大盤底部區間TB而進入一多頭走勢區間TBullish 後’可根據本發明方法建立的模糊邏輯規則庫,分析輸入 語意變數目前的技術指標間關係,以推論出目前應調整的 資金持股比例。 在詳細說明本發明的較佳實施例之後,熟悉該項技術 人士可清楚的瞭解,並在不脫離下述申請專利範圍與精神 下可進行各種變化與改變,而且本發明亦不受限於說明書 之實施例的實施方式,例如,本發明界定出大盤頭部區間 Ττ ’或者界定出大盤底部區間Tb可採用不同的日均線走 勢ΜΑΡ/Ό,如MAP/H)、或MAP㈣等,可使建立的模糊邏 輯規則庫獲得對短期、中期或長期的多頭走勢作出不同的 推論結果。 [發明功效] 經濟部智慧財產局員工消費合作社印製 根據本發明利用模糊邏輯分析股市多頭走勢資金持 股比例的方法所實施的人工智慧系統,可以從股市資訊的 歷史資料 PdGd)、Ρ\ν(Λν)、PmGm)、VD(/D)、Vw(iw)、Vm(,m)、 MAP/D(/D)與MAV/D(iD)中分析大盤指數的走勢關係,並由 加以界定出模糊邏輯規則庫,以提供投資者或操縱者在多 頭市場判斷目前技術指標間關係以推論出目前應調整的 資金持股比例。此外,本發明之知識資料庫或規則庫的實 施可以自動方式組織,而對當前股市多頭走勢實施適應性 模糊邏輯分析而使本發明方法提供了人工智慧解決的機 制。 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 A7 ____B7 五、發明説明() 23 綜上所述,本發明具有諸多優良特性,並解決習知技 術在實務上與應用上之缺失與不便,提出有效之解決方 法’完成實用可靠之系統,進而達成新穎且附經濟效益之 價値,實已符合發明專利之申請要件,懇請鈞局能予詳 審並賜准專利權益保障,以優惠民生實感德便。 [圖式之簡單說明] 圖一爲股票市場資訊之大盤指數與成交量的走勢圖。 圖二爲本發明方法中界定大盤指數之頭部區間的示 意圖。 圖三爲本發明方法中界定大盤指數之底部區間的示 意圖。 圖四爲本發明多頭走勢區間用以分割比例事件的時 段的時間分割比例圖。 圖五爲本發明實施例中語意變數之模糊子集合的歸 屬函數分布圖。 圖六爲本發明輸出語意變數之各比例的歸屬函 數圖。 圖七爲重疊模糊子集合合倂之歸屬函數圖。 圖八爲本發明資金持股比例方法之流程圖。 本紙張尺度適用中國國家標準(CNS ) A4規格(2丨〇><297公釐) I-----1!#! (請先閲讀背面之注意事項再填寫本頁) 訂 經濟部智慧財產局員工消費合作社印製573267 A7 B7 V. Description of the invention (π) The time division ratio shown in Figure 4 defines the nine proportional events E ,, / = 1 to 9 of the bull trend interval TBullish, among which, the fuzzy logic rule Rule_E / its semantic variables, k ^^ + DI > -DI, k ^^ D_IF > G's fuzzy sub-set € > D ,, the data sets are / ^ > d = {IQd⑺-Djd⑺}, / Gh dDi = {+ Dl6D ( 0 -DldD⑺}, C > 〇- {DIF ^ (〇}, tet] Therefore, the assignment function of each fuzzy sub-set 圮 &D; D, dDI, eF > is defined as follows: μ ^ 0 (Ι? Μ ^); I = min {K5d (〇-D6d (〇)? M = avg {K6D (〇-I > 6d (0)? R = max {K (5D (〇-D (5D (i))? Teti μ: οι >—D, (L, off; Z = min {+ DI6D (〇--DI6D (i)}, M = avg {+ DI6d⑺ 一 DI6d⑺), R = max {+ Dl6D (i) —Dl6D (i)} ? teti μ〇ΐΡ > 〇 (ΑΜ9τ?); L = minDIF5D (〇5 M = avgDIF ^ D (〇5 R = maxDIF (5D (〇5 tetj), where the function min is the smallest value in the data set; the function avg Is the arithmetic mean 资料 in the data set; the function max is the maximum 値 in the iei / data set. By analogy, the semantic variables zWexK of other fuzzy logic rules Rule_E, zWexK > D, " ito + D l > _Di, zWexDIF > The attribution function of the fuzzy sub-set corresponding to 0 household catties can also be defined in the above manner. Therefore, each long trend interval obtained from the historical data of the market's 6-day moving average map MAP6D (〇 Paper size applies Chinese National Standard (CNS) A4 specification (210 X 297 mm) (Please read the notes on the back before filling out this page) Order .. Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs and Consumer Cooperatives 573267 A7 B7 Printed by the Consumer Cooperative of the Property Bureau. V. Invention Description (18) TBullish can build a complete library of fuzzy logic rules, and input semantic variables can systematically define the attribution function of its fuzzy sub-set corresponding to technical index information. In addition, logic rules The attribution function of the semantic variables /? _ · 〇 of the result of the inference result of the fund's shareholding ratio of the library is defined as follows: L = 4, M = 9 μ: (I, M, 7?); L = |, M = 4 , R = 9 μ; 1 (Z, M, 7?); L = !, M = !, R = 4 Mw: (Dry, dry hole); 1, M = I, R = I ^ lim 5 M = 1? R = j μ: 1 (Ι? Μ? /?); L = |? M = |? R = 1ML) (厶, dry hole); > Shi, M = wu, R =. · Μ; 1 (Ι5Μ57?); L = i? M = i? R = | C (-, μ, acupoint); M = go, R = 士 and Figure 6 shows the assignment function of the proportions of the semantic variables of the inference results. Because the fuzzy logic rule base established from different long trend intervals TBullish in the historical data has a large number of fuzzy sub-sets organized by each input semantic variable, the present invention uses a combined operation to improve the efficiency of the fuzzy logic rule base inference. The overlapping fuzzy sub-sets are combined to reduce the number of fuzzy sub-sets of the input semantic variable partition. Please refer to Figure VII for a map of the assignment function of overlapping fuzzy sub-sets. The so-called "overlapping fuzzy sub-sets" refer to the assignment functions of these fuzzy sub-sets. This paper applies the Chinese National Standard (CNS) A4 specification ( 210 X 297 mm) (Please read the precautions on the back before filling out this page.) Pack. I-line · Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 573267 A7 B7 V. Description of the invention (19) A certain degree of overlap with each other ', and the combination operation system combines the two attribution functions into a new attribution function' to represent a new fuzzy sub-set, instead of the conditional combination of the two fuzzy sub-sets in the fuzzy logic rule. The operation is as follows: Among them, Li = min (Uy), M 丨, 7?] = Max ((footer, 矣) As for the condition of the combination of the two belonging functions, the parameters of the two belonging functions ^ fall within a set range. In one embodiment of the present invention, the setting range of the “attribution function combination” is determined according to the relationship between the technical indicators expressed, and the range is dynamically adjusted with the setting range. The number of cut fuzzy sub-sets will be effectively reduced. Therefore, after combining the fuzzy sub-sets of the input semantic variables, each fuzzy logic rule of the fuzzy logic rule base will be re-examined the conditional combination of its input semantic variables; In one embodiment of the present invention, 'If there are two or more fuzzy logic rules that infer the same capital shareholding result, the conditional combination of the fuzzy subsets corresponding to the input semantic variables is the same, then Equal fuzzy logic rules only need to retain one fuzzy logic rule in the rule base. In a further embodiment of the present invention, if there are two or more fuzzy logic rules inferring the same capital shareholding result, its input The conditional combinations of the fuzzy sub-sets corresponding to the semantic variables are not the same, then the conditional combinations of the fuzzy logic rules are logically ORed with each other to deduce the same capital shareholding result. This paper scale applies to China National Standard (CNS) A4 Specification (210 X 297 mm) (Please read the notes on the back before filling this page ) Assembly line 573267 A7 B7 V. Description of the invention (20) In a further embodiment of the present invention, if there are two or more fuzzy logic rules, the fuzzy subsets corresponding to the input semantic variables are combined with the same conditions When inferring different capital share ratio results, the attribution function that outputs semantic variables performs a merge operation to infer a new capital share ratio result. Therefore, the fuzzy logic rule base established according to the method of the present invention uses fuzzy Logic analysis fuzzifies the relationship between technical indicators of each semantic variable. After inference of each fuzzy logic rule, the inference results are defuzzified to obtain a capital shareholding ratio or The capital shareholding ratio 値; among them, the input technical indicator information is blurred by the attribution function of its fuzzy subset, and the inferred capital shareholding ratio is, for example, COA (Center of Area), COS (Center of Sums), and M0M (Mean of Maxima) defuzzification processing has been seen in the fuzzy logic analysis of conventional techniques. Finally, please refer to FIG. 8 for a flowchart of the method of capital holding ratio of the present invention. The method of the present invention uses fuzzy logic to analyze the long-term trend of the stock market to infer the proportion relationship between the funds held by investors and the funds held by the investor. The method includes the following procedures: Historical data on the stock market trend MAP / D (iD ) Defines the set of the head market interval Tτ, as shown in Figure 2, and the head market index curve Tτ is expressed by formula (5); defined in the historical data of the stock market trend MAP / D (iD) The collection of TB at the bottom of the market is shown in Figure 3. The paper size of the TB at the bottom of the market is applicable to China National Standard (CNS) A4 (210 X 297 mm) (Please read the precautions on the back before filling in this (Page) Order · -line · Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 573267 A7 B7 V. Description of the invention (21) The index curve is expressed by formula (8); (Please read the precautions on the back before filling in this (Page) The historical data of the stock market trend has been defined from the combination of the market's head interval Tτ and the market's bottom interval TB. The long trend interval of the market's bottom interval TB has been established ^ Bullish and expressed by formula (9); The time division ratio is set. In each bull trend interval TBumsh, the time period of the fund holding proportion event is defined. The conditional combination of the fuzzy logic rules for organizing the fund holding proportion event is defined by the technical indicator information. The combination of the relationship between the technical indicators reflected, and the fuzzy sub-set corresponding to the semantic variables of the relationship between the technical indicators in the specified period, the inference result of the fuzzy logic rule is the capital shareholding ratio; The fuzzy logic rule base of the fund holding proportion event organized in the long trend interval TBullish, the attribution function of its fuzzy subset is defined by the data of the relationship between the technical indicators corresponding to the semantic variables in the corresponding period; the wisdom of the Ministry of Economic Affairs The Property Cooperative Consumer Cooperative printed a combined operation to combine the fuzzy sub-sets of the semantic variables of the fuzzy logic rule base with a combined operation to reduce the number of fuzzy sub-sets divided by the semantic variables. The 倂 operation is expressed by equation (11). Further examining the fuzzy logic rule base, Conditional combination of fuzzy logic rules and inference results to remove fuzzy logic rules from the rule base that use the same combination of conditions to infer the result of the same capital shareholding ratio, reducing redundant or duplicate rules of the fuzzy logic rule base; Methods The fuzzy logic rule base was organized in an automatic way and the complexity of the rule base was reduced to provide investors with the current standard of the paper. The paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 public f " 573267 A7 ____ B7 V. Description of the invention (22) (Please read the notes on the back before filling in this page) After the market has established the bottom interval TB and entered a long trend interval TBullish ', a library of fuzzy logic rules that can be established according to the method of the present invention, Analyze the relationship between the current technical indicators of the input semantic variables to deduce the current shareholding ratio of funds that should be adjusted. After a detailed description of the preferred embodiment of the present invention, those skilled in the art can clearly understand, and can make various changes and modifications without departing from the scope and spirit of the patent application described below, and the present invention is not limited to the description For example, in the embodiment, for example, the present invention defines the market head interval Tτ 'or the market bottom interval Tb. Different daily moving average trends MAP / Ό, such as MAP / H), or MAP㈣ can be used to establish The fuzzy logic rule base obtains different inference results for short, medium or long bull movements. [Effects of the invention] An artificial intelligence system implemented by the method of analyzing the long-term stock holdings of funds in the stock market using fuzzy logic according to the present invention by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs can be obtained from historical data of stock market information (PdGd), P (Λν), PmGm), VD (/ D), Vw (iw), Vm (, m), MAP / D (/ D) and MAV / D (iD) analyze the trend relationship of the market index and define it A fuzzy logic rule base is provided to provide investors or operators to judge the current relationship between technical indicators in the bull market to deduce the current proportion of funds that should be adjusted. In addition, the implementation of the knowledge database or rule base of the present invention can be organized in an automatic manner, and adaptive fuzzy logic analysis is performed on the current bull market trend, so that the method of the present invention provides a mechanism for artificial intelligence to solve. This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 573267 A7 ____B7 V. Description of the invention () 23 In summary, the present invention has many excellent characteristics, and solves the problem that the conventional technology is practically compatible with The shortcomings and inconveniences in the application, put forward an effective solution 'complete a practical and reliable system, and then achieve a novel and economically valuable price, which has indeed met the application requirements for invention patents, I urge you to review and grant the patent rights Safeguarding the benefits of the people's livelihood. [Brief description of the diagram] Figure 1 is a trend chart of the broad market index and trading volume of stock market information. Figure 2 is a schematic diagram of the head zone defining the broad market index in the method of the present invention. Figure 3 is a schematic diagram illustrating the bottom interval of the broad market index in the method of the present invention. FIG. 4 is a time division ratio chart of a time period for dividing a proportional event by a bull trend interval of the present invention. Fig. 5 is an attribution function distribution diagram of fuzzy sub-sets of semantic variables in the embodiment of the present invention. Fig. 6 is a graph of the assignment function of each proportion of the output semantic variables of the present invention. Figure 7 shows the assignment function of the combination of overlapping fuzzy subsets. FIG. 8 is a flowchart of the method of capital holding ratio of the present invention. This paper size applies Chinese National Standard (CNS) A4 specification (2 丨 〇 > < 297 mm) I ----- 1! #! (Please read the notes on the back before filling this page) Order the Ministry of Economic Affairs Printed by the Intellectual Property Bureau Staff Consumer Cooperative

Claims (1)

573267 Α8 Β8 C8 D8 六、申請專利範圍 1.一種資訊分析方法,用於分析具有投資風險的資訊,尤其 是指股票市場的資訊,包含以下的程序: (請先閱讀背面之注意事項再填寫本頁) 在資訊的歷史資料MAP/D(/D)中界定出頭部區間Ττ 之第一集合; 在資訊的歷史資料MAP/D(iD)中界定出底部區間ΤΒ 之第二集合; 從前述第一集合與第二集合,組織多頭走勢區間 Teullish之弟二集合, 根據一預設的時間分割比例,界定前述第三集合中 每一多頭走勢區間TBullish內不同資金持股比例事件的時 段,藉以組織該資金持股比例事件所對應時段內的技術 指標間關係; 利用預設的歸屬函數界定每一多頭走勢區間 TBullish 內相對應時段中前述技術指標間關係的資料分佈;以及 線· 建立由模糊邏輯規則所組成的規則庫,其中前述模 糊邏輯規則的條件組合爲不同的前述技術指標間關係 之組合,而其推論結果爲前述資金持股比例。 經濟部智慧財產局員工消費合作社印製 2.如申請專利範圍第1項所述之資訊分析方法,其中前述 歷史資料包含大盤指數PD(iD)、Pw(iw)、PM(~)與成交量 VdGd)、Vw(iw)、VmGm) 0 3·如申請專利範圍第2項所述之資訊分析方法,其中前述 大盤指數Pd(〖d)的/日均線爲MAP/d(〖d)。 4.如申請專利範圍第1項所述之資訊分析方法,其中前述 界定頭部區間Ττ程序包含: 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 573267 A8 B8 C8 D8 六、申請專利範圍 a) 以大盤ί日均線MAPzD(iD)爲基礎,在時間軸化取 一時間區間T,其中,在時間區間T內,季線MAP72D 或半年線MAPdM或年線MAP·爲下凹曲線,且在該時 間區間T內,前述MAP/D(iD)具有至少一個極大値zm與 至少一個極小値zn ;以及 b) 決定一α値,以獲得前述頭部區間Ττ,使得MAP/d 存在 MAP/D(iD)>a ’ ,dgTt 且 MAP/D(iD)<a,茫丁丁。 5·如申請專利範圍第4項所述之資訊分析方法,其中前述 時間區間Τ可選自半年到一年之間。 6·如申請專利範圍第4項所述之資訊分析方法,其中前述 時間區間Τ可選自7個月到12個月之間。 7·如申請專利範圍第4項所述之資訊分析方法,其中前述 時間區間Τ最好選自30週到46週之間。 8·如申請專利範圍第4項所述之資訊分析方法,其中前述 MAP/D(,D)係選自以 MAP3D、MAP5D、^1八?;21:)或 ΜΑΡ· 的其中之一。 9·如申請專利範圍第4項所述之資訊分析方法,其中前述 頭部區間Ττ介於7日到21日之間爲佳。 10.如申請專利範圍第4項所述之資訊分析方法,其中前述 頭部區間Ττ最好是大約兩週左右。 11·如申請專利範圍第4項所述之資訊分析方法,其中前述 a値選自程序a)中極小値211之一。 I2.如申請專利範圍第1項所述之資訊分析方法,其中前述 界定底部區間TB之程序包含: 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) ' (請先閱讀背面之注意事項再填寫本頁) « . -·線· 經濟部智慧財產局員工消費合作社印製 573267 A8 B8 C8 D8 六、申請專利範圍 (請先閱讀背面之注意事項再填寫本頁) a) 以大盤/日均線走勢MAP/D(iD)爲基礎,在時間軸 化取一時間區間T,其中,在時間區間T內,季線MAP72D 或半年線map5M或年線map72M爲上凹曲線,且在該時 間區間T內,前述MAP/D(iD)具有至少一個極大値冗111與 至少一個極小値zn ;以及 b) 決定一β値,以獲得前述底部區間TB,使得MAP/d 存在 MAP/D(iD)Sp ’ /dgTb 且 ΜΑΡ/〇(/^)<β ’ 。 13. 如申請專利範圍第12項所述之資訊分析方法,其中前 述時間區間Τ可選自半年到一年之間。 14. 如申請專利範圍第12項所述之資訊分析方法,其中前 述時間區間Τ可選自7個月到12個月之間。 I5·如申請專利範圍第12項所述之資訊分析方法,其中前 述時間區間Τ最好選自30週到46週之間。 --線· I6.如申請專利範圍第I2項所述之資訊分析方法,其中前 述 MAP/D(rD)係選自以 MAP3D、MAP6D、MAP/2D 或 MAP2¥D的其中之一。 經濟部智慧財產局員工消費合作社印製 Π·如申請專利範圍第I2項所述之資訊分析方法,其中前 述底部區間ΤΒ介於7日到21日之間爲佳。 15. 如申請專利範圍第Π項所述之資訊分析方法,其中前 述底部區間ΤΒ最好是大約兩週左右。 19·如申請專利範圍第I2項所述之資訊分析方法,其中前 述β値選自程序a)中極大値2〇1之一。 2〇·如申請專利範圍第1項所述之資訊分析方法,其中前述 模糊邏輯規則的形式如下: 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) - 573267 A8 B8 C8 D8 夂、申請專利範圍 IF目前技術指標間關係是符合前述技術指標間關係之 組合的歸屬函數所界定的範圍, THEN目前投資資金與持股之間應爲前述資金持股比 例。 21·如申請專利範圍第1項所述之資訊分析方法,其中每一 則述多頭走勢區間TBullish係在確立前述底部區間下8後 到下一個前述頭部區間Ττ最大値之間。 22. 如申請專利範圍第i項所述之資訊分析方法,其中每一 前述多頭走勢區間TBullish係在確立前述底部區間丁8後 所形成的走勢。 23. 如申請專利範圍第1項所述之資訊分析方法,其中前述 預設的時間分割比例在前述多頭走勢區間TBullish的初 段到尾段間則述資金持股比例的持股所佔比例遞減。 24. 如申請專利範圍第1項所述之資訊分析方法,其中前述 預設的時間分割比例在前述多頭走勢區間TBullish的初 段到尾段間前述資金持股比例的資金所佔比例增加。 25·如申請專利範圍第1項所述之資訊分析方法,其中前述 技術指標間關係的技術指標包含隨機指標(KD値)、平 滑異同平均線(MACD)、相對強弱指標(RSI)、心理線 (PSY)、動量指標(MTM)、乖離率(BIA)、騰落指標 (ADL)、漲跌比率(ADR)、逆勢操作(CDP)、振盪指標 (OSC)等技術指標。 26.如申請專利範圍第25項所述之資訊分析方法,其中前 述技術指標間關係之組合爲K値扣値、+DIF>-DIF或D 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) 訂· 線· 經濟部智慧財產局員工消費合作社印製 573267 A8 B8 C8 D8 六、申請專利範圍 値>80之全部或部分組合 (請先閱讀背面之注意事項再填寫本頁) 訂· 線· 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐)573267 Α8 Β8 C8 D8 6. Scope of Patent Application 1. An information analysis method used to analyze information with investment risks, especially information on the stock market, including the following procedures: (Please read the precautions on the back before filling in this Page) defines the first set of head intervals ττ in the historical data MAP / D (/ D) of the information; defines the second set of bottom intervals TB in the historical data MAP / D (iD) of the information; The first set and the second set, organize the second set of the long trend interval Teullish, according to a preset time division ratio, define the time period of the event of different funds holding proportions in each long trend interval TBullish in the aforementioned third set, thereby Organize the relationship between the technical indicators in the period corresponding to the capital shareholding event; use the preset attribution function to define the data distribution of the relationship between the aforementioned technical indicators in the corresponding period in each bull trend interval TBullish; and A rule base composed of fuzzy logic rules, wherein the conditions of the aforementioned fuzzy logic rules are combined into different aforementioned technologies The relationship between a combination of standard, and its corollary result of the aforementioned capital stake. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 2. The information analysis method described in item 1 of the scope of patent application, where the aforementioned historical data includes the broad market index PD (iD), Pw (iw), PM (~) and transaction volume VdGd), Vw (iw), VmGm) 0 3. The information analysis method as described in item 2 of the scope of patent application, wherein the daily average of the broad market index Pd (〖d) is MAP / d (〖d). 4. The information analysis method described in item 1 of the scope of patent application, wherein the aforementioned procedure for defining the head section Tτ includes: This paper size is applicable to the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 573267 A8 B8 C8 D8 6. Scope of Patent Application a) Based on the broader market daily average line MAPzD (iD), take a time interval T on the time axis, where within the time interval T, the seasonal line MAP72D or the semi-annual line MAPdM or the annual line MAP is A concave curve, and within the time interval T, the aforementioned MAP / D (iD) has at least one maximum 値 zm and at least one minimum 値 zn; and b) determines an α 値 to obtain the aforementioned head interval ττ so that the MAP / d MAP / D (iD) > a ', dgTt and MAP / D (iD) < a, Mantintin. 5. The information analysis method according to item 4 of the scope of patent application, wherein the aforementioned time interval T can be selected from half a year to one year. 6. The information analysis method according to item 4 of the scope of patent application, wherein the aforementioned time interval T can be selected from 7 months to 12 months. 7. The information analysis method according to item 4 of the scope of patent application, wherein the aforementioned time interval T is preferably selected from 30 weeks to 46 weeks. 8. The information analysis method according to item 4 of the scope of patent application, wherein the aforementioned MAP / D (, D) is selected from the group consisting of MAP3D, MAP5D, and ^ 1? ; 21 :) or one of MAP. 9. The information analysis method as described in item 4 of the scope of patent application, wherein the aforementioned header interval Tτ is preferably between 7th and 21st. 10. The information analysis method according to item 4 of the scope of patent application, wherein the aforementioned head interval Tτ is preferably about two weeks. 11. The information analysis method described in item 4 of the scope of patent application, wherein the aforementioned a 値 is selected from one of the minimum 値 211 in the program a). I2. The information analysis method as described in item 1 of the scope of patent application, wherein the aforementioned procedure for defining the bottom interval TB includes: This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) '(Please read first Note on the back, please fill out this page) «.-· Line · Printed by the Consumers' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 573267 A8 B8 C8 D8 6. Scope of patent application (please read the notes on the back before filling this page) a) Based on the market / day moving average trend MAP / D (iD), take a time interval T in the time axis. Among them, within the time interval T, the seasonal line MAP72D or the semi-annual line map5M or the annual line map72M is a concave curve, and Within the time interval T, the aforementioned MAP / D (iD) has at least one maximum redundancy 111 and at least one minimum 値 zn; and b) determines a β 値 to obtain the aforementioned bottom interval TB, so that MAP / d exists in MAP / d D (iD) Sp '/ dgTb and MAP / 〇 (/ ^) < β'. 13. The information analysis method according to item 12 of the scope of patent application, wherein the aforementioned time interval T can be selected from half a year to one year. 14. The information analysis method according to item 12 of the scope of patent application, wherein the aforementioned time interval T can be selected from 7 months to 12 months. I5. The information analysis method according to item 12 of the scope of patent application, wherein the aforementioned time interval T is preferably selected from 30 weeks to 46 weeks. -Line · I6. The information analysis method described in item I2 of the scope of patent application, wherein the aforementioned MAP / D (rD) is selected from one of MAP3D, MAP6D, MAP / 2D or MAP2 ¥ D. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Π. The information analysis method described in item I2 of the scope of patent application, where the aforementioned bottom interval TB is preferably between 7th and 21st. 15. The information analysis method described in item Π of the patent application scope, wherein the bottom interval TB is preferably about two weeks. 19. The information analysis method described in item I2 of the scope of patent application, wherein the aforementioned β 値 is selected from one of the largest 値 2 in the program a). 20. The information analysis method described in item 1 of the scope of the patent application, wherein the form of the aforementioned fuzzy logic rule is as follows: This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm)-573267 A8 B8 C8 D8 范围 Scope of patent application IF The current relationship between technical indicators is in the range defined by the attribution function of the combination of the foregoing technical indicators. The current investment funds and shareholdings of THEN should be the shareholding ratios of the aforementioned funds. 21. The information analysis method described in item 1 of the scope of patent application, wherein each of the long trend intervals TBullish is between the time when the bottom interval is established 8 and the next head interval Tτ maximum 値. 22. The information analysis method described in item i of the scope of patent application, wherein each of the aforementioned long trend intervals TBullish is a trend formed after the aforementioned bottom interval D8 is established. 23. The information analysis method described in item 1 of the scope of patent application, wherein the preset time division ratio decreases from the beginning to the end of the aforementioned bullish trend interval TBullish. 24. The information analysis method as described in item 1 of the scope of patent application, wherein the preset time division ratio increases between the proportion of funds held by the aforementioned capital shareholding ratio from the beginning to the end of the aforementioned bullish trend interval TBullish. 25. The information analysis method as described in item 1 of the scope of patent application, wherein the technical indicators of the aforementioned technical indicators include a stochastic indicator (KD 値), a smooth similarity average (MACD), a relative strength index (RSI), and a psychological line (PSY), momentum indicator (MTM), deviation rate (BIA), take-off indicator (ADL), up-down ratio (ADR), contrarian operation (CDP), oscillator index (OSC) and other technical indicators. 26. The information analysis method as described in item 25 of the scope of patent application, wherein the combination of the aforementioned technical indicators is K 値 buckle, + DIF > -DIF or D. The paper size applies the Chinese National Standard (CNS) A4 specification ( 210 X 297 mm) (Please read the notes on the back before filling out this page) Order · Thread · Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economy 573267 A8 B8 C8 D8 Part of the combination (please read the notes on the back before filling this page) Order · Thread · Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs, Consumer Cooperatives This paper is sized to the Chinese National Standard (CNS) A4 (210 X 297 mm)
TW91107036A 2002-04-09 2002-04-09 Fuzzy-based index-relation AI method for holding ratio in bullish market TW573267B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW91107036A TW573267B (en) 2002-04-09 2002-04-09 Fuzzy-based index-relation AI method for holding ratio in bullish market

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW91107036A TW573267B (en) 2002-04-09 2002-04-09 Fuzzy-based index-relation AI method for holding ratio in bullish market

Publications (1)

Publication Number Publication Date
TW573267B true TW573267B (en) 2004-01-21

Family

ID=32734123

Family Applications (1)

Application Number Title Priority Date Filing Date
TW91107036A TW573267B (en) 2002-04-09 2002-04-09 Fuzzy-based index-relation AI method for holding ratio in bullish market

Country Status (1)

Country Link
TW (1) TW573267B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI488138B (en) * 2012-10-12 2015-06-11 Mitake Information Corp Device and method of providing automatic technical analysis of stock market

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI488138B (en) * 2012-10-12 2015-06-11 Mitake Information Corp Device and method of providing automatic technical analysis of stock market

Similar Documents

Publication Publication Date Title
Dunis Artifi cial Intelligence in Financial Markets
Naranjo et al. A fuzzy decision system for money investment in stock markets based on fuzzy candlesticks pattern recognition
Brunnermeier et al. The maturity rat race
Dempster et al. An automated FX trading system using adaptive reinforcement learning
Thu et al. Using support vector machine in FoRex predicting
Biswas et al. Stock Market Prediction: A Survey and Evaluation
Cohen Artificial Intelligence in Trading the Financial Markets
Han et al. ROBO-ADVISORS: REVOLUTIONIZING WEALTH MANAGEMENT THROUGH THE INTEGRATION OF BIG DATA AND ARTIFICIAL INTELLIGENCE IN ALGORITHMIC TRADING STRATEGIES
Guo et al. Market Making with Deep Reinforcement Learning from Limit Order Books
Jia et al. Machine learning of surrender: Optimality and humanity
TW573267B (en) Fuzzy-based index-relation AI method for holding ratio in bullish market
Alamili Exchange rate prediction using support vector machines
Kovalerchuk et al. Interpretable Machine Learning forFinancial Applications
Solares et al. A Comprehensive Soft Computing-Based Approach to Portfolio Management by Discarding Undesirable Stocks
Peachavanish Dual time frame relative strength stock selection using fuzzy logic
TW518493B (en) Fuzzy-based methods for holding ratio in bearish market
TW591458B (en) Fuzzy-based methods for holding ratio in bullish market
TWI226993B (en) Fuzzy-based methods for holding ratio in stock market
Chatziloizos et al. Forecasting Stock Market Trends using Deep Learning on Financial and Textual Data.
Zhu et al. Dual-market quantitative trading: The dynamics of liquidity and turnover in financial markets
Nosratabadi et al. Credit risk assessment of bank customers using DEMATEL and fuzzy expert system
Yu et al. FinMe: A Performance-Enhanced Large Language Model Trading Agent with Layered Memory and Character Design
Lyukevich et al. Cryptocurrency Market: Choice of Technical Indicators in Trading Strategies of Individual Investors
Wilkman Feasibility of a reinforcement learning based stock trader
Nikulin et al. Fuzzy portfolio optimization model with estimation of results

Legal Events

Date Code Title Description
GD4A Issue of patent certificate for granted invention patent