個人資料
正文

An old ox pulling a rickety cart---making slow progress

(2010-08-06 03:32:39) 下一個
GOOGLE TRANSLATION

A cow\'s technical analysis course (ZT)
Source: gutu

Phase 1 is to learn some traditional technical analysis indicators, such as MACD, KDJ, RSI, etc.. Found significant uncertainty.

Phase 2 study programmed with flying fox, download a lot of people to the indicators, but also learned Vb, found significant uncertainty, although the pattern infinite, but the essence of the traditional technical indicators and there is no difference, are established historical data on the basis of a simple average just all.

3 stage, the pursuit of more powerful statistical analysis, learning the SPASS, playing familiar with time series analysis of ARIMA. Found significant uncertainty. ARIMA residuals of the original white noise is not estimated.

Stage 4, Learning GARCH, damn SPASS actually do not have this tool, only to learn MATLAB7. GARCH playing cooked, we found lot of uncertainty. Original, GARCH is still essentially linear estimation, but will continue to ARIMA ARIMA residuals once. Fainted.

Stage 5, the network N by some people 忽悠 artificial neural network, started to play BP, RBF, found great uncertainty. BP, RBF fitting historical data is simply perfect, but the generalization of the future, simply is. Still did not give up, but also to improve the genetic algorithm crunching, using the chaos theory of phase space improvements, is still dog feces.

6 stage, listening to an artificial intelligence expert, said Nanda, SVM is the most NB, and continue to learn, this thing is difficult, at last, or to the buttoned up, the results found that significant uncertainty. Correct rate was disappointing.

Phase 7, the occasion of a loss, but also say N, is said to be useful wavelet, get looking through the book, feeling extremely difficult. But this time on the technical analysis has been shaken. Met a friend one day, combat expert, I talk to a presentation, discovery, by the wars, or some old-fashioned traditional indicators that the best is to use the magic key to a.

8 stage, at this stage, re-playing that traditional indicators of those few old-fashioned.
-------------------------------------------------- ----------
gutu see posts blind evaluation :

The cattle were beef cattle in comparison to hard, how much to be versatile, have learned circle, in the jar show show, the stage has been put up pos. Light is no longer useless conclusion : technical analysis, technical analysis is belated. But he still has not started, or did not find a their own, can make their way to make money with confidence. He only met a play some old-fashioned traditional indicators of the actual master, and re-learn and understand that a few old-fashioned indicators. He may also encounter a few will be different then the actual combat martial arts expert, I do not know how the situation would be?

In short, he had not realized that regular activity performed to set their own way. If the back is long to say, stop there! D D : :

一個牛人的技術分析曆程(ZT)
來源: gutu 於 09-09-04 19:19:56 [檔案] [博客] [舊帖] [轉至博客] [給我悄悄話]





一個牛人的技術分析曆程(ZT)

第1階段,是學習一些傳統的技術分析指標,如MACD,KDJ,RSI 等等。發現不確定性很大。

第2階段,學習用飛狐編程序,下載個許多人編製的指標,還學習了Vb,發現不確定性很大,雖然花樣無窮,但本質上與傳統的技術分析指標沒有差別,都是建立在對曆史數據簡單的各種均線基礎上而已。

第3階段,追求更厲害的統計分析,學習了SPASS,玩熟了時間序列分析ARIMA。發現不確定性很大。原來ARIMA對白噪音的殘差沒有估計。

第4階段,學習GARCH,該死的SPASS居然沒有這個工具,隻好學習MATLAB7。GARCH玩熟後,發現不確定性很大。原來,GARCH本質上依然是線性估計,不過是將ARIMA的殘差繼續ARIMA了一次。暈倒。

第5階段,被一些網絡N人忽悠人工神經網絡,開始玩BP,RBF,發現不確定性很大。BP,RBF對曆史數據的擬合簡直是完美,但對未來的泛化,簡直是。仍然不死心,又搗鼓用遺傳算法改進,用混沌理論的相空間改進,依然是狗屎。

第6階段,聽南大的一個人工智能專家說,SVM是目前最NB的,繼續學習,這玩意很難,終於還是給搞定了,結果,發現不確定性很大。正確率讓人失望。

第7階段,茫然之際,又有N人說,據說小波可能有用,找來書翻翻,感覺無比艱深。而此時對技術分析已經信心動搖。某日遇一朋友,實戰高手,一席交談演示,發現,靠,實戰中還是傳統的那幾個老掉牙的指標最好,關鍵是是運用之妙了。

第8階段,目前階段,重新玩那傳統的那幾個老掉牙的指標。
------------------------------------------------------------
gutu看帖瞎評:

這個牛人牛就牛在比較刻苦,多少“十八般武藝”都學了一圈,在壇子裏show花架子,擺pos的階段已經過了。已經不再輕下結論:技術分析無用,技術分析是馬後炮。但是他還是沒有入門,還是沒有找到一個屬於自己的,可以讓自己有把握地掙錢的方法。他隻遇到了一個“玩傳統的幾個老掉牙的指標”的實戰高手,而重新學習與體會那“幾個老掉牙”的指標。他可能後來還會遇到幾個會不同武功的實戰高手,不知情況又會如何?

一句話,他還沒有悟到,做股票要有一套屬於自己的方法。後麵的話再說就長了,就此打住! :D :D




[ 打印 ]
閱讀 ()評論 (1)
評論
目前還沒有任何評論
登錄後才可評論.