Predictable variation and profitable trading of US equities: a trading simulation using neural networks

Surveying stock market forecasting techniques – Part II: Soft computing methods

“…Studies that utilize neural networks include Constantinou et al (2006) who uses a MLP Network with 2 inputs, Motiwalla and Wahab (2000), who uses a back propagation neural network and Refenes, Azeme-Barac, and Zapranis (1993) who utilizes a feed forward network with 4 layers (two hidden ones). Two special cases are probabilistic neural networks (PNN), a standardized architecture neural network used by Kim and Han (1998) and the radial basis function network (RBFN) that has two layers and it is a special class of multilayer feed-forward networks; each unit in the hidden layer employs a radial basis function, such as a Gaussian kernel, as the activation function.…”

“…yield, vol., price/earnings ratio Kim and Han (1998) Technical analysis factors Kimoto et al (1990) Vector curve, turnover, interest rate, foreign exchange rate, Dow Jones index Kosaka et al (1991) Prices of three hundred stocks Koulouriotis (2004) Stock trend, stock profit, market profit, supply, demand Koulouriotis et al (2005) Market trend/prof., demand & supply, forces, P-days ahead price change Kuo (1998) Technical analysis factors Lam (2001) Twelve market indicators Leigh et al (2002) Twenty two technical analysis factors Lendasse et al (2000) Twenty five technical analysis variables Malliaris and Salchenberger (1993) Exer., days, close price, volume, int., lag close price, lag mark. price Mizuno et al (1998) Eleven technical indicators of TOPIX Motiwalla and Wahab (2000) Twenty technical analysis variables Nishina and Hagiwara (1997) Ten input of daily data Olson and Mossman (2003) Sixty one accounting and financial ratios Pai and Lin (2005) Daily stock data Pan et al (2005) Last six daily closing prices Pantazopoulos et al (1998) Daily closing value of index Perez-Rodriguez et al (2004) Daily stock data Phua et al (2001) Volume, opening, lowest, hhighest and closing index price, Dow Jones index value, NASDAQ index value, HIS index value, NIKKEI index value Qi (1999) Nine financial and economic variables Quah and Srinivasan (1999) Economical, political and firm/stock specific factors Raposo et al (2002) Five fundamental analysis indicators Rast (1999) Daily closing price of index Rech (2002) Daily closing price of index Refenes et al (1993) Three Technical Analysis factors Safer and Wilamowski (1999) Four price ratio averages, four volume ratio averages, one previous SUE Schumann and Lohrbach (1993) Thirteen economic time series Setnes and Van Drempt (1999) AEX price and macroeconomic factors Schumann and Lohrbach (1993) Shortrate, USD, DJ, Bonds, MSeuro Simutis …”

“…Lam (2001) Genetic algorithm Leigh et al (2002) HIT, t-Test Lendasse et al (2000) HIT Malliaris and Salchenberger (1993) MAD, MAPE, MSE Mizuno et al (1998) HIT Motiwalla and Wahab (2000) Eleven statistical indices for direct accuracy Oh and Kim (2002) APE, RMSE, MAE Olson and Mossman (2003) HIT, RETURN Pai and Lin (2005) MSE Pan et al (2005) RMSE, VR, HIT Pantazopoulos et al (1998) RMSE, HIT Qi (1999) RMSE, MAE, MAPE, PCC Quah and Srinivasan (1999) HIT Raposo et al (2002) HIT Rast (1999) HIT Rech (2002) RMSE, MAE, Diebold-Mariano Statistic Refenes et al (1993) RMSE, POCID, DIRECTION Safer and Wilamowski (1999) MSE Schumann and Lohrbach (1993) HIT Setnes and Van Drempt (1999) RMSE, SIGN%, WEALTH% (Siekmann et al, 1999) HIT, RMSE, PROFIT Situngkir and Surya (2003) Auto correlation Steiner and Wittkemper (1997) RETURN Tabrizi et al (2000) NMSE, learning rate Tan et al (1995) HIT, r, false alarm rate Thawornwong and Enke (2004) RMSE Tsaih et al (1998) HIT Vanstone et al (2005) Return. SD, PROFIT, DOF, Zscore Versace et al (2004) t-test, x 2 -test Wah and Qian (2002) HIT, MSE Wikowska (1995) MSE Wittkemper and Steiner (1996) MSE Wu et al (2001) HIT Yumlu et al (2004) MSE, HIT, TIC, correlation Yumlu et al (2005) MAE, RMSE, correlation Zhang et al (2004) HIT Zorin and Borisov (2002) RMSE…”

Strategy Modeling and Classifier Training for Share Trading

“…Various input parameters have also been considered by researchers to forecast stock returns, such as stock prices [8], trading volume, dividend, yield, indexes [7], technical indicators [9], and technical analysis variables [10]. In our approach, we use the history of the stock price and the signals generated by the user-defined strategies.…”

Combining seasonal time series ARIMA method and neural networks with genetic algorithms for predicting the production value of the mechanical industry in Taiwan

“…Luvai et al [7] posited that neural networks provide several advantages over regression prediction techniques. These advantages include neural networks require almost no assumptions regarding the underlying data to be forecasted and being able to develop models from incomplete or imperfect data.…”