1248007 九、發明說明： 【發明所屬之技術領域】 本發明係關於一種資訊預測方法，更明確的說，本發明是一種實 施於一人工智慧股票趨勢預測系統中，對於股市資訊的變化進行追縱 分析與預測以評估市場交易的方法。 【先前技術】 依先前技術，評選具有投資價值的股票可以透過下列幾個方法： 一、基本分析1248007 IX. Description of the invention: [Technical field of invention] The present invention relates to an information prediction method. More specifically, the present invention is implemented in an artificial intelligence stock trend prediction system to track changes in stock market information. Analysis and forecasting to assess market trading methods. [Prior Art] According to the prior art, the following methods can be used to select stocks with investment value: 1. Basic analysis
基本分析法即是考慮會影響價格的各種因素來預測市場未來 走勢的方法。投資人必須觀察分析各種可能對市場價格產生影響 的政治、經濟等因素，並試圖從商品價格背後的供給與需求關係 來預測未來價格變動的趨勢，如以投資一家公司為例其作法如下： 1·找出欲購買股票公司沿革與產業經濟背景。 2·估算該股淨值盈餘與股利。 3. 4. 5. 估算該股產銷情形與營運展望及財務概況。The basic analysis method is a method to predict the future trend of the market by considering various factors that affect the price. Investors must observe and analyze various political and economic factors that may affect market prices, and try to predict the trend of future price changes from the relationship between supply and demand behind commodity prices. For example, investing in a company is as follows: 1 · Identify the evolution of the company and the industrial economy. 2. Estimate the net surplus and dividend of the stock. 3. 4. 5. Estimate the production and sales situation and operational outlook and financial profile of the stock.
轉該股未㈣絲性，本舰衫合理，發放股利政 策，並與市場上其它同類型標的作比較。 根據以上各項數據或朗，即可作域資者騎投資股 票之依據。當錄本分析不僅要考慮上述因素，亦需考 慮其它外在_素，例如：政治因料，必須綜合種種 因素最後才能蚊。在找尋轉市場_素時，由於可 ^會遺漏其它重要因素，灿在騎基持析時，投資 广具有相關的財經知識’才能分析出結果，這對於 司tr技:者而s ’是—個很大的障礙。-般而言’公 …目關資訊’通常會落後市場反應，如果投資者等到 5 1248007 該公司公佈資訊後，才進行分析預測，通常市場上已經 反應完畢，結果可能會造成追高殺低的情況，甚至經濟 體系結構改變或非預期之變化，都可能導致市場供需失 衡，影響價格之預測。 二、技術面分析 技術为析主要是利用圖形或量化技術指標來研判市場商品 交易之人氣或供需之情況，進而決定買賣的時機。由於技術分析 不僅可以預測價袼並研判市場投資的時機，而且可作為預警之訊 號更重要的疋較基本分析簡單易懂，所以廣為一般大眾所使 用，以下將介紹在市場上常用到的技術分析之方法： 1 ·道氏理論（Dow Theory) 道氏理論是建立在趨勢（Trends)和形態走勢 (Patterns)之上，趨勢是由支撐線與壓力線構成，而形 態走勢則分成整理形態（Continuation Patterns)與反轉 形態（Reversal Patterns)兩種。今天大眾所熟知的多頭 市場（Bullish Market)、空頭市場（Bearish Market)、支 撐（Support)、壓力（Resistance)、頭肩形（Head andThe transfer of the stock is not (four) silky, the jersey is reasonable, the dividend policy is issued, and compared with other similar types on the market. According to the above data or lang, it can be used as the basis for the domain investors to ride investment stocks. When the book analysis needs to consider not only the above factors, but also other external factors, such as political factors, it is necessary to combine various factors to finally obtain mosquitoes. In the search for the market _ prime time, because it can omit other important factors, when the investment in the base is held, the investment has a wide range of relevant financial knowledge to analyze the results, which is for the division of the technology: s ' is - A big obstacle. - Generally speaking, 'public...directed information' usually lags behind the market reaction. If the investor waits until 5 1248007, the company publishes the information before analyzing and forecasting. Usually, the market has already completed the reaction, and the result may result in chasing high and low killing. Situations, even changes in economic structure or unanticipated changes, may lead to imbalances in market supply and demand, affecting price forecasts. Second, the technical analysis of technology is mainly based on the use of graphical or quantitative technical indicators to determine the market commodity trading popularity or supply and demand, and then decide the timing of trading. Because technical analysis can not only predict the price and determine the timing of market investment, but also as a warning signal, it is more important. It is easier to understand than basic analysis, so it is widely used by the general public. The following will introduce the technologies commonly used in the market. Method of analysis: 1 · Dow Theory The Dow Theory is based on Trends and Patterns. The trend is composed of support lines and pressure lines, while the shape trend is divided into sorting patterns ( Continuation Patterns) and Reversal Patterns. Bullish Market, Bearish Market, Support, Resistance, Head and Shoulder (Head and Head)
Shoulders Formation)等等，皆是由道氏理論發展出 來。了解道氏理論分析方法後，再結合統計學，結果衍 生出現今多種的價量技術分析方法。 2.波浪理論（Wave Theory) 由艾略特（R· N. Elliott )於1943年提出，此理論 認為市場價格是具有週期性與波動性，一個完整循環走 勢主要有八波，包括五個主波段和三個回標波，這個理 論結合了趨勢線、目標區預測及自然定律，是現今市場 3 常用的分析工具之一。 動平均線（Moving Average) 年多動平均線（ΜΑ)是最受廣泛使用的技術分析工 具’主要原因是其淺顯易懂、容易計算。由於它是以統 °十方法計算出一段時間内股票平均價格，再將各點連接 JLx 紙一條平滑的曲線，所以移動平均線具有平滑性、穩定Shoulders Formation and so on, all developed by Dow Theory. After understanding the Dow's theoretical analysis method, combined with statistics, the results are derived from a variety of price analysis techniques. 2. Wave Theory (Wave Theory) was proposed by R. N. Elliott in 1943. This theory holds that market prices are cyclical and volatility. There are eight waves in a complete cycle, including five mains. The band and three reticle waves, combined with trend lines, target area predictions, and natural laws, are among the three commonly used analytical tools in today's market. Moving Average The annual moving average (ΜΑ) is the most widely used technical analysis tool. The main reason is that it is easy to understand and easy to calculate. Since it calculates the average price of the stock over a period of time in a unified method, and then connects each point to a smooth curve of JLx paper, the moving average is smooth and stable.
It及趨勢性等特徵。移動平均線是用來決定買賣時機的 參考’美國投資專家Joseph Granville提出的八大法 則’就是針對移動平均線的買賣時機而訂定的。 4· 成交量（V〇L) 利用成交量移動平均線可判斷趨勢向上或向下。當 成交量增加時，即代表趨勢向上，而成交量減少時，即 代表趨勢向下。當股票要上漲時，一定要配合成交量， 右股價上旅但成交量卻縮小，產生價量背離的情況，則 股價可能只是短期反彈。注意，當成交量上漲時，卻不 一定保證股價一定會上漲，需配合其它因素與技術指標 才能研判漲或跌。Features such as It and trend. The moving average is the reference used to determine the timing of buying and selling. The eight rules proposed by US investment expert Joseph Granville are set for the timing of buying and selling the moving average. 4. Volume (V〇L) Use the volume moving average to determine whether the trend is up or down. When the volume increases, it means that the trend is upward, and when the volume is reduced, it means the trend is downward. When stocks are going to rise, they must match the volume. If the right stock price goes up but the trading volume shrinks, and the price is deviated, the stock price may only be a short-term rebound. Note that when the volume of trading increases, it does not necessarily guarantee that the stock price will rise. It is necessary to cooperate with other factors and technical indicators to judge whether it will rise or fall.
5· RSI RSI係利用某段時間内的股價變動情形，預測未來 價格的變動趨勢。其基本原理為在一個正常股市中多空 買賣雙方的力道，必須均衡的，股價才會穩定。而RSI 是計算在一定期間内，股價上漲總幅度平均值佔總涨跌 幅總幅度平均值的比例。以6日RSI值為例，其值若 為80以上為超買，90以上或Μ頭為賣點；20以下 為超賣’ 10以下或W底為買點；RSI會比股價變動 1248007 先出現峰或底，能預先反映股價的漲跌趨勢，可視為大 盤指數走勢的先行指標；若當6日RSI由下往上穿過 12日RSI時，可視為買點；反之，當6日RSI由 上在下貫破12曰RSI時’可視為賣點；RSI線走勢 與大盤指數走勢呈背離現象代表大盤即將反轉。5. RSI RSI uses the stock price changes over a period of time to predict future price movements. The basic principle is that in a normal stock market, the strength of both buyers and sellers must be balanced and the stock price will be stable. RSI is the ratio of the average value of the stock price increase to the average of the total increase and decrease in a certain period of time. Take the 6-day RSI as an example. If the value is 80 or more, it is overbought, 90 or more or Shantou is the selling point; 20 or less is oversold '10 or below or W is the buying point; RSI will peak before the stock price changes 1248007 or At the end, it can reflect the ups and downs of the stock price in advance, which can be regarded as the leading indicator of the trend of the market index; if the RSI on the 6th passes through the 12-day RSI from the bottom up, it can be regarded as the buying point; otherwise, when the RSI is on the 6th, the RSI is from the top to the bottom. When the 12-inch RSI is broken, it can be regarded as a selling point; the trend of the RSI line and the trend of the large-cap index indicate that the market is about to reverse.
6. KD 將RSI強弱指標、移動平均線以及量能觀念的優 點予以融合之技術指標。如果行情是一個明顯的漲勢， 會帶動K線與D線向上升，如漲勢開始遲緩，則會 反應到K值與D值，使得κ值跌破D值，此時中 短期跌勢確立；當K線自上往下跌破D線，且D值 在8〇以上（超買區），為賣出訊號；當κ線自下向上 突破D線，且值在2〇以下（超賣區）出現，為買進 訊號；當K值大於80，D值大於7〇時，表示當日 收盤價處於偏尚之價格區域，即為超買狀態；當K值 小於20，D值小於30時，表示當日收盤價處於偏低6. KD combines the advantages of RSI strength indicators, moving averages and quantitative energy concepts. If the market is a clear uptrend, it will drive the K and D lines to rise. If the rally starts to be sluggish, it will reflect the K and D values, so that the K value falls below the D value, and the short-term downtrend is established. When the K line falls from the top to the broken D line, and the D value is above 8〇 (overbought area), it is the sell signal; when the κ line breaks through the D line from the bottom up, and the value is below 2〇 (oversold) Area) appears as a buy signal; when the K value is greater than 80 and the D value is greater than 7〇, it means that the closing price of the day is in the price range, which is the overbought condition; when the K value is less than 20 and the D value is less than 30, Indicates that the closing price of the day is at a low level
之價格區域，即為超賣狀態；價格創新高或新低，而KD 未有此現象，此為背離現象，亦即為可能反轉的重要前 兆。 在股票市場中有眾多的資訊要分析，#然相_技術指標種類也 繁多，在此僅列舉上述幾個較為常用之技術指標分析。其它尚有日線、 週線、月線、融資、融卷..·等資訊，投f者要如何處理如此多的資訊 呢？雖践腦可以很容易計算分析各種技標，但單純的技術指標 分析會有落《勢與指標鈍化的情況發生，且沒有自鮮習、判斷修 正的能力，所以其分析預測之結果非常不準確。希望透過本發明可使 1248007 電腦具有人工智慧（AI)，如此一來，不僅可以計算、具有自我學習修 正的能力，還可適應股市各種情況的變動，進而能更準確地預測股市 未來的趨勢。 在分析預測的賴巾，有眾多的技術且各有所長，但具有自我修 正與學習的分析預測方法中，帛為人所熟知的便是類神經網路（版㈤ ne_k)與模糊（Fuzzy)邏輯推論。對於股票市場而言，很多的資 訊具有不確定性與波祕，軸兩者亦可做出不錯的效果，但類神經 網路需要大量足_資訊才可得到結果，且對股市變動迅速的特性而 言=無法很快學f及修正，且計算成本較高。賴㈣輯推論在定義 其歸屬函數（Membership funetion )並不容易，且其規則亦很難去訂定。 而灰色關連分析適驗灰色系統，對於部分信息已知，部分未知的系 統稱為灰色系統，灰色關連分析法計算簡單且剌面廣，例如可應用 在危險事故、保險、農工業、商業、經濟、金融商品…等之分析，其 不僅具有可適祕股票市場之特性，且只f要少數資訊即可建立其模 型進行分析預測，亦不需要去定義歸屬函數或訂定相關規則。八、 在股票市場中，對-般投資者而言，幾乎都是大賠小賺，導致資 金常被套牢以致週轉不靈或被迫斷頭等等。一般投資者在買賣股票 時，在往為規避風險而分散投資標的，但又無法同時並長期的追縱掌 握自己所投資的標的，或缺乏風險控管的能力，導致投資損失，但若 將標的集中單-個股，則其投資風險也相對升高，可能_次就將所有 籌碼的都輸掉。 從前述的說明可以瞭解，在股票市場中有太多的技術指標方法存 在，同時要瞭解股票之基本面，例如財務狀況、未來成長性等，還要 本握整個大環境之經濟面、政治面等因素，這對一般的投資者而言， 實在是難以掌握。股票市場常充斥著各種流言與小道消息，往往造成 1248007 投資者的困擾，進而影響投資者做出正確的判斷，種種因素使得一般 投資者心中對此形成一道難以跨越之鴻溝。 在股票市場中’趨勢的判斷是最重要的，只要能掌握趨勢就能掌 握獲利的關鍵。股票市場不外乎「多頭」、「空頭」、「盤整」三種狀態， 若預測未來股票市場為「多頭」走勢，則投資者可以進場買股票增加 持股比例。反之，若預測未來股票市場為「空頭」走勢，則投資者可 以儘速獲利了結出場或融卷放空，若預測未來股票市場為「盤整」則 應避免大篁進出並降低持股比例’甚至退出市場觀望，以降低投資風 險。投資者若可以掌握未來股票市場趨勢，則投資股票獲利的機會將 大增而投資之風險也將隨之降低。 前述之股票投資方法，不管是使用技術分析或基本分析都是落後 指標，雖然仍有其用處，但是往往緩不濟急，反而造成投資者追高殺 低，成為有心人士坑殺的對象。此外，影響投資者判斷的，還包含各 種經常出現在股市的消息面，以及投資者自己的情緒問題，這往往造 成負面的判斷。有鑑於此，所以本發明發展出一具有人工智慧的方法， 並經由實施本發明之系統，使電腦成為具有「人的智慧」的人工智慧 系統。此一系統不僅可從所接收的資訊實際反應預測結果，可不受消 息面與個人情緒影響，還可取代以「人」為決策中心的問題，達到以 「人智」代替「人治」的目的。 【發明内容】 為了具有人工智慧的能力，可以幫投資者做預測及決策，本發明 採用了灰色理論中的灰關連分析的方法。對於部分信息已知，部分未 知的系統可稱之為灰色系統，在股票市場中有很多的訊息是未知的’ 所以股票市場為一灰色系統，非常適合用灰色理論的應用。 1248007 斑預、=方:發:的主要目的係提供-種對股市資訊的變化 即通知㈣者未來事件（f、賣或平倉）的發生 幫助技_貝者易於投資，並降低投資風險與提昇投資效 進^亍追縱 以評估市 ^發明的另—目的係提供_種基於股票市場中判斷趨勢 ~父易的方法。 本發明的再-目的係提供一種在股票市場中基於領先市場的指 欠在月·』述資料中可以瞭解股票市場中相關資訊與指標繁多，一般投 資者根本無法同時且長期追蹤所有的訊息變化，並且容易受外在環境 及個人情緒影響，造成雜錯誤與投f失利。本發鶴提供—種資訊 追蹤預測的方法，對大盤指數變化進行追蹤與賴，並人工智慧 =方法，預測大盤未來的狀態為「多頭」、「空頭」或「盤整」，以供投 資者可立即做出正柄投㈣斷與行為（wf或平倉），而不再需要 花費大量時間精力去研究各種相關的資料及報告。 在股票市場中有很多的訊息是未知的，所以股票市場是一個灰色 系統，適合用灰色理論的方法。在本發明較佳實施例中，首先以過去 股票市場中大盤指數的歷史資料作為原始輸入資料，再來可分成三部 分，其中前二個部分是利用技術指標分析方法來判斷大盤現在狀態為 該買或賣，至於另外一個部分，則是利用灰關聯分析方法來判斷大盤 現在狀態是「多頭」、「空頭」或「盤整」訊息。接著，結合三者之分 析結果，以預測未來大盤的趨勢狀態為「多頭」、「空頭」或「盤整」， 最後再辅以風險控管機制，決定投資人應該採取「買進」、「賣出」或 「平倉」之操作方式。 11 1248007 【實施方式】 以下本發明將對較佳實施例及所附之圖予以充分描述，但在此描 述之前應瞭解熟悉本行之人士可修改在本文中描述之實施例創作，同 時獲致本發明之同等功效。因此，須瞭解以下之描述對熟悉該項技藝 之人士而言為一廣泛之揭示，且其内容不在於限制本發明。以下為與 本發明背景有關之技術的延伸描述。雖然有嫻熟經驗及知識的讀者可 選擇僅跳讀或甚至不讀以下之背景資訊，但了解此項資訊可進一步掌 握本發明的具體實施，建議應加以詳讀。 首先參考圖一所示，為習知灰色關聯分析的流程圖，其詳細過程 如下列之步驟所示： 1·確定參考數列與比較數列，將參考數列（％(&))與比較數 列UW)因子確定，其中AAeZ且ζ· = 〇，ι，2，...，„， 工。(灸)，义(^:)分別表示Ά在（點的數值。 2·在確認參考數列與比較數列後，需將數列作前處理，也就 是前一小節所提灰關聯生成的方法，使參考數列與比較數 列之物理意義或度量單位相同。 3·經過前處理後，即可求出〜對乂的灰關聯係數（Grey Relational Coefficient，GRC)，其公式如式 U 所示，其中， ζ為分辨係數(Distinguishing Coefficient)，ζ€[〇，ι],通常取 ζ = 0·5，△〇，·(々）因所採用的模型而有所不同計算方式。 4·求出灰關聯係數後，即可求出數列間的灰關聯度 Relational Grade，GRG)。X是所有信息序列 ' 的集合， ζ· = 0，1，2,…，W取其中你）分別為^ 在k點的數值。若滿足公式12，°則 丫(乂，\)稱為〜對X。的灰關聯度。 12 1248007 另參考圖二所示，為本發明方法之流程圖。在本發明方法之實施 例中，首先從股票市場中大盤指數3、櫃臺指數及電子類股的歷史資 料作為原始數據21，在圖三中所示為大盤κ線圖，依其時間軸不同 而有日線、週線、月線之分。在日κ線圖中，每筆資料以一日為基 礎而週、月Κ線圖，則分別以一週與一月為基礎。在本實施例中 輸入資料時，則是以採用大盤日Κ線圖之日均線為輸入的資訊，例 如3日均線（MAp3D)、6日均線（MAP6D)、24日均線（MAP24D) 等，本發明可分成兩個部分來加以討論，其中一部份為技術指標22、 23分析。在本發明較佳實施例中，在眾多技術指標中，取移動平均價 (MAP)、移動平均量（MAV)、移動平均融資（MAF)為原始輸入數據， 同時採用多種不同的平均線MAp、MAV、MAF分別判斷大盤屬於多 頭、空頭或盤整。如果股市行情是一個明顯的漲勢，則天數較小的平 均線會排在天數較大的平均線之上，屬於多頭市場，為一買入訊號； 當天數較小的平均線排在天數較大的平均線之下，意味市場為空頭市 % ’為一賣出訊號。在本發明較佳實施例中，MAp取1㊉3、3㊉72、 20㊉72、24㊉144與3G㊉72五組均線組合；MAV取5φ144、6φ72、 6㊉144、10㊉12與12㊉15五組均線組合；MAF取j㊉72、3φ72、 5㊉72、6㊉72與1〇㊉72五組均線組合作為判斷條件，其中㊉表示 長期與短期均線交又比較。 在技術指標（CBF) 22分析中（詳細流程請參考圖四），選取大盤 (Tfx)，檯（0TC)與電子類股（ELE)，將其主要因素（成交價、成 交量及融資）分別作職。成交價進行MAp條件測試，成交量進行 MAV條件測試，融資則進行MAF條件測試，其中MAp、mav、 MAF泰式條件為上述之結果，最後將三個類股執行、 MAF⑷籍件後之值各自乘上其權值加總，並將錢股洲買、賣訊 13 1248007 號之值做相加，即為技術指標（CBF) 22分析之輪出結果。 在技術指標（PBF) 23分析（詳細流程請參考圖五）中，其基本 原理與CBF #法相同。技術指標22 /分析是針對每一種類股的成交 價、成交量_資進行研妹m㈣的思考料。技翻標2s 分析方法，則是-種橫向的思考方式，是針對每—種要素（成交價、 成交量與融資）進行分析。每次執行時，只選取-X貞要純行測試流 程，將各類股執行_罐之條件值做加總，然後將加總後之結果做臨 界值測試。當所有要素皆完成臨界值測試後，將所有同意買、賣或盤 整訊號做相加，而相加之結果即為技術指標23分析輸出之結果，最 後將技術指標22、23分析之結果當成風險控管機制25之輸入。 另外一個部分為灰色關聯分析24 (詳細流程請參考圖六），即技 術指標（GBF) 24分析。在原始數據21中，本發明之較佳實施例， 分別對大盤（Taiex)、櫃檯（〇TC)與電子類股（ELE)之成交價 (MAP)做測試，較佳實施例中Taiex取ΜΑ3—28、·5一27、 MA10一 14、MA12一32、MA72一 18 五條均線；〇TC 取 MA3 50、MA10 19、 — — MA12一 16、MA20一30、MA144一27 五條均線；ELE 取 MA5_14、 MA12一 16、MA24一35、MA72一 17、MA144—34 五條均線。上述 MAx y 中的x表示x天均線，y表示時間區間長度，而時間區間從5〜5〇 天0 再利用前述灰色關聯分析方法，決定參考數列與比較數列，參考 數列是利用道氏理論與移動平均線結合之特性推出，為一固定之函數 會根據比較數列之情況做相對應之修正，決定參考數列與比較數列 後，即可進行多、空線型比對。計算GRG的方法有很多種，如圖一 中之程序11、12。在本發明中各種計算方式之結果雖然有所差異，但 是其次序（Order)則不變，所以本發明直接採用距離GRG之計算方 1248007 式’即圖-之作法。當鋪股所有比較數列皆執行完成，重複上述作 法’執行其它的類股，再將所有同意w、f或盤整訊號做相加，即為 灰色關聯分析24輸出之結果，最後亦將其結果t成風險㈣機制25 之輸入。 經過技術指標分析22、23與灰色關聯分析24三種方法測試 後，可各自得到的-組買、賣或盤整狀態（BuyDe㈣、s·嗯ee、 HoldDe㈣）之結果，當成風隨錢制25讀人（詳細流程請參 考圖七）。將各組結果輸入其相對應方法的風險控管函數中，其中各分 析方法風險控管函數皆是由實驗過程中求得，將f、f或盤整狀態結 果針對原始系統產生之買、賣點進行分析比較，#原始操作訊號產生 買（賣）點時，則判斷賣（買）點狀態之風險函數，求出其相對應之 風險程度。 經過風險控管函數處理後，將三種分析方法產生之結果再做加權 處理，將具有較佳效果之權值提高，再把三種方法加權後之結果相加 (SafeDegree)，最後將SafeDegree結果，依風險分類函數將買、賣點 分成高、中、低風險三類輸出，即為本發明最終顯示預測之結果26。 如圖三所示為大盤K線圖，依其時間軸不同而有日線、週線、 月線之分。在本實施例中皆採用日K線圖的日均線為輸入資料，並 分別利用短期與長期之均線組合作為預測之判斷條件。當然投資者所 觀察之均線上亦可採用其它曰、週或月均線等用以判斷指數之趨勢， 其判斷方法亦如同圖二之流程所示，只是在程序22、23與24選取 之測試條件有所差異而已。 在本發明較佳實施例中，選取大盤（Taiex)、櫃檯（〇TC)與電子 類股（ELE)作為原始數據之來源，所有分析皆是針對此三項做處理。 投資者可將其改為其它類股，例如金融、證券等類股，或在分析時加 15 1248007 必須另外經由 上其它之類股，只是不同類股之判斷條件會有所不同 實驗求出，而其後續作法流程，則完全相同。 另在本發雜佳實施射，將縣要素__ 前述的流程一樣 W、或在分析時加上其它之要素，作為測試之條件，其餘之做法與The price area is oversold; the price is high or low, and KD does not have this phenomenon. This is a divergence phenomenon, which is an important precursor to possible reversal. There are a lot of information to be analyzed in the stock market, and there are many types of technical indicators. Here are just a few of the more commonly used technical indicators. Others still have daily, weekly, monthly, financing, financing, etc., and how do you deal with so much information? Although it is easy to calculate and analyze various technical indicators, it is easy to analyze and analyze the technical indicators. However, the situation of the potential and the indicator passivation occurs, and there is no ability to self-study and judge the correction. Therefore, the results of the analysis and prediction are very inaccurate. . It is hoped that the 1248007 computer can have artificial intelligence (AI) through the invention, so that it can not only calculate, have the ability of self-learning and correction, but also adapt to changes in various conditions of the stock market, thereby more accurately predicting the future trend of the stock market. In the analysis and prediction of the towel, there are many techniques and their own strengths, but in the analysis and prediction methods with self-correction and learning, what is well known is the neural network (version (f) ne_k) and fuzzy (Fuzzy). Logical inference. For the stock market, a lot of information has uncertainty and volatility, and the axis can also make good results, but the neural network needs a lot of information to get the results, and the characteristics of the stock market change rapidly. For example, = can't learn f and fix quickly, and the calculation cost is higher. Lai (4) series inference is not easy to define its membership function (Membership funetion), and its rules are difficult to set. The gray correlation analysis is suitable for the gray system. For some information, some unknown systems are called gray systems. The gray correlation analysis method is simple and wide-ranging, for example, it can be applied to dangerous accidents, insurance, agro-industry, commerce, economy. The analysis of financial products, etc., not only has the characteristics of the stock market, but only a small amount of information can be used to establish its model for analysis and prediction, and there is no need to define the attribution function or to set relevant rules. Eight, in the stock market, for the average investor, almost all of the big losers make small profits, resulting in the funds are often stuck so that the turnover is not working or forced to break the head and so on. When investors buy and sell stocks, they are diversifying their investment targets to avoid risks. However, they cannot simultaneously and long-termly grasp the target of their own investment, or lack the ability to control risks, resulting in investment losses, but if the target is If you concentrate on single-shares, your investment risk will increase relatively, and you may lose all your chips in _ times. From the above description, we can understand that there are too many technical indicators in the stock market, and at the same time, we must understand the fundamentals of stocks, such as financial status, future growth, etc., but also the economic and political aspects of the entire environment. Such factors, for the average investor, are really difficult to grasp. The stock market is often filled with all kinds of rumors and gossip, which often causes 1248007 investors to trouble, which in turn affects investors to make correct judgments. Various factors make the general investor's mind form an insurmountable gap. The judgment of the trend in the stock market is the most important, as long as you can grasp the trend, you can grasp the key to profit. The stock market is nothing more than three states: “long”, “short” and “consolidated”. If the future stock market is predicted to be “long”, investors can enter the market to buy stocks to increase their shareholding ratio. Conversely, if the future stock market is predicted to be a “short” trend, investors can profit as soon as possible to make a profit or a short-selling. If the future stock market is forecast to “consolidate”, it should avoid entering and lowering the shareholding ratio even Exit the market and wait and see to reduce investment risks. If investors can grasp the future stock market trends, the chances of investing in stock profits will increase and the risk of investment will also decrease. The aforementioned stock investment method, whether using technical analysis or basic analysis, is a backward indicator. Although it still has its usefulness, it is often unsatisfactory. Instead, it causes investors to chase after high and kill, and become the target of people who are interested in killing. In addition, the influence of investors' judgments also includes various news situations that often appear in the stock market, as well as investors' own emotional problems, which often lead to negative judgments. In view of this, the present invention develops a method with artificial intelligence, and by means of the system embodying the invention, the computer becomes an artificial intelligence system with "human intelligence". This system can not only reflect the predicted results from the received information, but also be free from the influence of the message surface and personal emotions. It can also replace the problem of "people" as the decision-making center and achieve the purpose of replacing "people's governance" with "human intelligence". SUMMARY OF THE INVENTION In order to have the ability of artificial intelligence, investors can make predictions and decisions. The present invention adopts the method of gray correlation analysis in the grey theory. For some information, some unknown systems can be called gray systems, and there are many messages in the stock market that are unknown. So the stock market is a gray system, which is very suitable for the application of gray theory. 1248007 Spot Pre-, = Fang: Send: The main purpose is to provide a kind of change to the stock market information, that is, to notify (4) the future event (f, sell or close) of the help technology _ shellers easy to invest, and reduce investment risks and Improve the efficiency of investment into the 亍 亍 縱 縱 縱 縱 縱 縱 縱 縱 縱 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ The re-purpose of the present invention provides a kind of information and indicators in the stock market that can be understood based on the leading market in the stock market. The general investors cannot track all the information changes at the same time and in the long term. And is easily affected by the external environment and personal emotions, causing misunderstandings and failures. The company provides a method for tracking and forecasting information, tracking and changing the index of the market, and artificial intelligence = method to predict the future status of the market as "long", "short" or "consolidation" for investors. Immediately make a positive handle (four) break and behavior (wf or close), and no longer need to spend a lot of time to study various related information and reports. There are a lot of messages in the stock market that are unknown, so the stock market is a grey system that is suitable for gray theory. In the preferred embodiment of the present invention, firstly, the historical data of the market index in the past stock market is used as the original input data, and then can be divided into three parts, wherein the first two parts use the technical index analysis method to judge the current state of the market. Buying or selling, as for the other part, is to use the gray correlation analysis method to judge whether the current status of the market is "long", "short" or "consolidated". Then, combined with the analysis results of the three, to predict the future trend of the broader market as "long", "short" or "consolidation", and finally with the risk control mechanism, the investors should adopt "buy" and "sell" The operation mode of "out" or "close". DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT The present invention will be fully described in the following description of the preferred embodiments and the accompanying drawings, but it should be understood that those skilled in the art can modify the embodiments described herein and obtain the present invention. The same effect of the invention. Therefore, it is to be understood that the following description is a broad disclosure of those skilled in the art and is not intended to limit the invention. The following is an extended description of the technology related to the background of the present invention. Although readers with extensive experience and knowledge may choose to skip only or not read the following background information, understanding this information will further exemplify the specific implementation of the present invention and the recommendations should be read in detail. Referring first to Figure 1, the flowchart of the conventional gray correlation analysis is as follows. The detailed process is as follows: 1. Determine the reference sequence and the comparison sequence, and compare the reference sequence (%(&)) with the comparison sequence UW. The factor is determined, where AAZ and ζ· = 〇, ι, 2, ..., „, work. (moxibustion), meaning (^:) respectively indicate Ά (value of the point. 2. Confirmation of reference series and comparison After the series, the series needs to be pre-processed, that is, the method of generating the gray relation in the previous section, so that the reference number and the comparison series have the same physical meaning or unit of measurement. 3. After pre-processing, it can be found ~ Relation Gray Relational Coefficient (GRC), the formula is shown by the formula U, where ζ is the Distinguishing Coefficient, 〇€[〇, ι], usually ζ = 0·5, △〇 ,·(々) is calculated differently depending on the model used. 4. After finding the gray correlation coefficient, the gray relational degree between the series can be obtained. Relational Grade, GRG). X is the set of all information sequences. , ζ· = 0,1,2,...,W take you among them) respectively ^ at k If the formula 12 is satisfied, then 丫(乂,\) is called the gray correlation degree of ~ to X. 12 1248007 Referring additionally to Figure 2, a flow chart of the method of the present invention is implemented. In the example, the historical data of the market index 3, the counter index and the electronic stocks in the stock market is taken as the raw data 21, and the large-scale κ line graph is shown in Figure 3. The daily and weekly lines are different according to the time axis. In the daily κ line chart, each piece of data is based on one day, and the weekly and monthly line charts are based on one week and one month respectively. In the case of inputting data in this embodiment, The present invention can be divided into two parts, such as a 3-day moving average (MAp3D), a 6-day moving average (MAP6D), a 24-day moving average (MAP24D), etc., in which information is input using a daily average of the daily chart. Part of the analysis is technical indicators 22, 23. In the preferred embodiment of the invention, among the many technical indicators, the moving average price (MAP), the moving average (MAV), and the moving average financing (MAF) are taken as the original input. Data, using a variety of different average lines MAp, MAV, MAF Don't judge whether the market is long, short or consolidating. If the stock market is a clear trend, the average number of days will be ranked above the larger average, which is a long market, a buy signal; The smaller average line is ranked below the larger average line, meaning that the market is a short market %' is a sell signal. In a preferred embodiment of the invention, MAp takes 1 3 3, 3 10 72, 20 10 72, 24 10 144 and 3 G 10 72 Five groups of moving averages; MAV takes 5φ144, 6φ72, 6: 144, 10:12 and 12:15 five-group moving average combination; MAF takes j, 72, 3φ72, 5:72, 6:72, and 1〇10,72 sets of moving averages as the judgment condition, of which ten means long-term and short-term moving average Compare again. In the analysis of technical indicators (CBF) 22 (refer to Figure 4 for detailed process), select the main market (Tfx), Taiwan (0TC) and electronic stocks (ELE), respectively, the main factors (transaction price, volume and financing) Work. The transaction price is tested by MAp condition, the transaction volume is tested by MAV condition, and the financing is tested by MAF condition. The MAp, mav and MAF Thai conditions are the above results. Finally, the three stocks are executed, and the values of MAF(4) are the respective values. Multiply the sum of its weights and add the value of Qianshouzhou Buy and Sell 13 1248007, which is the result of the analysis of technical indicators (CBF). In the Technical Indicators (PBF) 23 analysis (see Figure 5 for the detailed process), the basic principle is the same as the CBF # method. Technical Indicator 22 / Analysis is based on the consideration of the transaction price and volume of each type of stock. The technical analysis of 2s analysis method is a horizontal way of thinking, which is to analyze each element (transaction price, volume and financing). For each execution, only -X is selected to be a pure-line test process, and the condition values of various types of stock execution _ cans are summed, and then the summed results are tested for critical value. After all the factors have completed the critical value test, all the agreed buy, sell or consolidation signals are added, and the result of the addition is the result of the analysis and output of the technical indicator 23, and finally the result of the analysis of the technical indicators 22, 23 is regarded as the risk. The input of the control mechanism 25 is used. The other part is the gray correlation analysis 24 (see Figure 6 for the detailed process), which is the technical indicator (GBF) 24 analysis. In the raw data 21, the preferred embodiment of the present invention tests the transaction price (MAP) of the Taiex, the counter (〇TC) and the electronic stock (ELE), respectively. In the preferred embodiment, Taiex takes 3 —28,·5·27, MA10-14, MA12-32, MA72-18, five moving averages; 〇TC takes MA3 50, MA10 19, —MA12-16, MA20-30, MA144-27 25 moving average; ELE takes MA5_14 , MA12-16, MA24-35, MA72-17, MA144-34 five moving averages. The x in the above MAx y represents the x-day moving average, y represents the length of the time interval, and the time interval is from 5 to 5 days. The gray correlation analysis method is used to determine the reference sequence and the comparison sequence. The reference sequence is based on the Dow theory. The combination of moving averages is introduced. For a fixed function, the corresponding corrections are made according to the comparison sequence. After determining the reference sequence and the comparison sequence, multiple and empty line comparisons can be performed. There are many ways to calculate GRG, as shown in Figure 11, program 11. Although the results of various calculation methods in the present invention are different, but the order is unchanged, the present invention directly adopts the method of calculating the formula of the GRG by the method of the GRG. When all the comparison series of the stocks are executed, repeat the above method 'execute other stocks, and then add all the agreed w, f or consolidation signals, which is the result of the gray correlation analysis 24 output, and finally the result is t Risk (IV) Input of Mechanism 25. After the technical indicators analysis 22, 23 and the gray correlation analysis 24 three methods test, each can get the results of the group buy, sell or consolidation (BuyDe (four), s hm ee, HoldDe (four)), when the wind with the money system 25 readers (Please refer to Figure 7 for the detailed process). The results of each group are input into the risk control function of the corresponding method. The risk control functions of each analysis method are obtained from the experiment process, and the f, f or consolidation state results are made for the buying and selling points generated by the original system. Analysis and comparison, when the #original operation signal generates a buy (sell) point, it judges the risk function of the selling (buying) point state, and finds the corresponding risk level. After the risk control function is processed, the results of the three analysis methods are weighted, the weights with better effects are increased, and the weighted results of the three methods are added (SafeDegree), and finally the SafeDegree results are The risk classification function divides the buying and selling points into high, medium and low risk three types of outputs, which is the result of the final display prediction for the present invention26. As shown in Figure 3, the large-scale K-line chart has daily, weekly, and monthly lines depending on its time axis. In the present embodiment, the daily average line of the daily K-line chart is used as the input data, and the combination of the short-term and long-term moving averages is used as the judgment condition of the prediction. Of course, the average line observed by investors can also use other 曰, weekly or monthly moving averages to judge the trend of the index. The judgment method is also shown in the flow of Figure 2, but the test conditions selected in the procedures 22, 23 and 24. There are differences. In the preferred embodiment of the invention, the Taiex, the counter (〇TC) and the electronic stock (ELE) are selected as the source of the raw data, and all analyses are performed for these three items. Investors can change it to other stocks, such as financial, securities and other stocks, or add 15 1248007 in the analysis must be through other stocks, but the judgment conditions of different stocks will be different experimentally. The follow-up process is exactly the same. In addition, in the implementation of this issue, the county element __ the same process as W, or other factors in the analysis, as a test condition, the rest of the practice and
在本發明較佳實施射，其效果不論是在測試資料㈣㈣D 料（85.1G〜91.2)，皆優於不採用本發明之做法，因此本發明 ' ^、賣訊號做更進—步分析’蚊最後買賣點之風險程度，提 供投資者一個具有低風險且高獲利的投資方法。 又 本發明方法的流財，程序22、23為技術減之分析，在本發 明之f佳纽财以MAP、MAV、MAF粒__件。當股市 灯情是-個乡财場，則天數較小的平均線排在天數較大的平均線之 ^，此為買入訊號；當天數較小的平均線排在天數較大的平均線之下， 意味市場為空頭市場，此為賣出訊號。在程序22、23中，因為預測 大盤指數趨勢是中短期的而非長期的，且股票市場的變動非常的迅 速’所以均線或其它技術指標之時__取不能太長，若太長則無 法迅速反應股市實際狀況，但亦*能太短，否則可紐軸之波動所 欺騙，所以本發明皆選取五組均線組合作為篩選條件，而均線之長度 則從3〜144天。 '、 同理，在本發明之方法流程中，程序24為灰色關聯分析，亦選 取五條均線作為篩選條件，均線之長度從3〜144天，時間之長度則 從5〜60天，在本發明實施例中篩選條件皆選取五組作為訓練條件， 投資者亦可選取更多或更少之訓練條件。 從上述說明可知道，股票市場有眾多的技術指標分析方法，例如·· 1248007 、成交量，融資、融卷、移動平均線等。在此可改 m kd、macd，或加人其它技術指標於實施例之 中，而其整個預測判斷之流程，除了 2 23因不同技術指標， 有不同條件觸式外，其它之程序皆同前述之方法。 另外在程序24巾，即灰色關聯分析法，本發明是採用計算距離 GRG之A式w然，亦可採用斜率、面積等計算之公式，其 預測結果與_ GRG之以_騎果，軸會有所差異，但其^欠 序_er)貝|J不變。這意謂著，其它作法步驟完全與灰色關聯分析計 算距離GRG作法相同。 最後在程序25 +，將程序22、23與24之結果，做進一步筛 選並導入風陳管概念，求出最後W、f之風陳度，_由程序％ 顯示判斷結果’並通知投資者未來股市漲跌聰，以避免投資風險太 高。例如，若預測大盤會上張，但卻顯示其風險高，則本發明提供相 關訊息通知投資者，降低投資風險，達到風險控管目的。 在詳細說明本發明之作法流程後，可以清楚瞭解，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 於說明書之實施方式。例如，大盤指數可改為個股之股價、大盤指數 可改為大盤成交量，或技術指標MAP、MAV、MAF可改為KD、MACD 荨方式。 根據本發明之人工智慧系統，可將經嚴格篩選後的股市之歷史資 料：大盤（Taiex)、櫃檯（OTC)與電子類股（ELE)作為原始數據之 來源，以不同的MAP、MAV與MAF等均線，經由人工智慧的方法， 透過CBF、PBF與GBF等三種技術指標分析與風險控管的處理，以 預測出未來股市之趨勢，除了可立即通知投資者未來股市趨勢的狀態 外，還可提高投資者的穫利，並有效降低風險與損失。 17 1248007 本發明之實施方法已詳述於前述 領域之人士皆可依本發明之說明，在齡 #何熟悉本技術 内視需要更動、修飾本發明，因此=本發明之精神與範圍 明之申請專利範圍中。 、施態樣亦包含在本發 特性，不僅可解決一些實 綜合以上所述，本發明實具有諸多優良 際應用上的缺失與不便，提歧濟有效 ^ :響’進而有效控管投資的風險，實已符合發明專利之巾請要件，懇 0月狗局能予詳審並賜予專利權保障，以優惠民生實感德便。 1248007 【圖式簡單說明】 圖一為習知灰色關聯分析之流程圖。 圖二為本發明分析預測方法之流程圖。 圖三為大盤κ線圖。 圖四為本發明技術指標（CBF)之流程圖。 圖五為本發明技術指標（PBF)之流程圖。 圖六為本發明灰色關聯分析（GBF)之流程圖。 圖七為本發明風險控管之流程圖。 【主要元件符號對照說明】 大盤指數 3 灰關聯係數公式 11 灰關聯度公式 12 原始數據程序 21 技術指標程序（CBF) 22 技術指標程序（PBF) 23 灰色關聯分析（GBF) 24 風險控管機制程序 25 顯示預測結果程序 26 大盤指數i日均線走勢MAPiD 31 3曰均線MAP3D 32 19In the preferred embodiment of the present invention, the effect is better than the non-invention of the present invention in the test data (4) (4) D material (85.1G~91.2), so the present invention '^, sell the signal to do more step-by-step analysis' mosquito The degree of risk of the last trading point provides investors with a low-risk and highly profitable investment method. Further, the method of the present invention is carried out, and the programs 22 and 23 are technically reduced. In the present invention, the 佳, 纽, and MAF particles are __ pieces. When the stock market is a township, the average number of days is ranked on the larger average of the days, which is the buy signal; the smaller average is ranked on the larger average. Underneath, it means that the market is a short market, this is a sell signal. In the procedures 22, 23, because the forecast of the market index trend is short-term and not long-term, and the stock market changes very quickly 'so the average or other technical indicators __ can not be too long, if it is too long, it can not Quickly reflect the actual situation of the stock market, but it can also be too short, otherwise it can be deceived by the fluctuation of the new axis. Therefore, the present invention selects five groups of moving average combinations as the screening conditions, and the length of the moving average is from 3 to 144 days. Similarly, in the method flow of the present invention, the program 24 is a gray correlation analysis, and five moving averages are also selected as the screening conditions, the length of the moving average is from 3 to 144 days, and the length of the time is from 5 to 60 days, in the present invention. In the examples, the screening conditions are selected as five training conditions, and investors can also select more or less training conditions. From the above description, we can know that there are many technical indicators analysis methods in the stock market, such as ··································· Here, m kd, macd, or other technical indicators can be changed in the embodiment, and the entire process of predicting and judging, except for 2 23 due to different technical indicators, different conditional touches, other procedures are the same as the foregoing The method. In addition, in the program 24 towel, that is, the gray correlation analysis method, the present invention adopts the formula A of calculating the distance GRG, and can also calculate the formula of the slope, the area, etc., and the prediction result and the _ GRG are _ riding fruit, the axis will There is a difference, but its ^ is out of order _er) Bay|J is unchanged. This means that the other method steps are exactly the same as the gray correlation analysis calculation distance GRG. Finally, in the program 25 +, the results of the programs 22, 23 and 24 are further filtered and introduced into the concept of wind and pipe, and the final W and f winds are determined, and the result is displayed by the program % and the investor is notified. In the future, the stock market will go up and down to avoid too high investment risks. For example, if the forecast is to be released, but it shows that the risk is high, the present invention provides relevant information to notify investors to reduce investment risks and achieve risk control purposes. It will be apparent that various changes and modifications can be made without departing from the scope and spirit of the invention. For example, the market index can be changed to the stock price of individual stocks, the market index can be changed to the market volume, or the technical indicators MAP, MAV, MAF can be changed to KD, MACD. According to the artificial intelligence system of the present invention, the historical data of the strictly selected stock market: Taiex, counter (OTC) and electronic stocks (ELE) can be used as the source of the original data, with different MAP, MAV and MAF. The average moving line, through the artificial intelligence method, through the analysis of three technical indicators such as CBF, PBF and GBF and the management of risk control, in order to predict the trend of the future stock market, in addition to immediately informing investors of the state of the future stock market trend, Increase investor profitability and effectively reduce risks and losses. 17 1248007 The method of the present invention has been described in detail in the above-mentioned fields, and the invention can be modified and modified according to the needs of the present invention. Therefore, the patent application of the spirit and scope of the present invention is as follows. In the scope. The mode of application is also included in the characteristics of this issue, which can not only solve some of the above-mentioned actual synthesis, but also has many shortcomings and inconveniences in the application of the invention, and it is effective to improve the risk of investment. It has already met the requirements for the invention patent towel. In January, the dog bureau can give a detailed examination and grant the patent right to protect the people's livelihood. 1248007 [Simple description of the diagram] Figure 1 is a flow chart of the conventional gray correlation analysis. Figure 2 is a flow chart of the analysis and prediction method of the present invention. Figure 3 shows the κ line chart of the market. Figure 4 is a flow chart of the technical indicator (CBF) of the present invention. Figure 5 is a flow chart of the technical indicator (PBF) of the present invention. Figure 6 is a flow chart of the Gray Correlation Analysis (GBF) of the present invention. Figure 7 is a flow chart of the risk control of the present invention. [Main component symbol comparison description] Large market index 3 Grey correlation coefficient formula 11 Grey correlation degree formula 12 Raw data program 21 Technical index program (CBF) 22 Technical index program (PBF) 23 Grey correlation analysis (GBF) 24 Risk control mechanism program 25 Display forecast results program 26 Market index i daily average trend MAPiD 31 3曰 moving average MAP3D 32 19