Skip to main content

Table 1 Comparison of related studies

From: A comprehensive evaluation of ensemble learning for stock-market prediction

ArticlesBase learnerNumber of weak learnersEnsemble algorithms (combination method)DatasetsMachine learning taskEvaluation methods
[19]MLP BAGTokyo stock exchangeStock  
[18]SVM and NNNot statedSTKS&P 500 indexStock Cross-validation
[20]RNNNot statedSTKAmerican Association of Individual InvestorsStock Accuracy
[12]LR, NN, K-NN and SVM BAG, BOTAmadeus databaseStock Operating characteristic curve (AUC)
[23]SVM10MVSao Paulo Stock Exchange IndexStock 10-fold CV
[21]DT, RF and ANNNot statedBAG, BOTU.S stock (S&P 500)Stock MAPE, RMSE, R2
[22]RF30BAGNASDAQ  Accuracy, precision
Recall and specificity
[24]BMA WALS and LASSO BOT-BAGNot statedstock Out-of-sample R2
[28]NN1–5BAGMexican stock exchangeStock  
[37]SVM and extra tress1–250STKNot statedStock RMSE
[25]TressNot statedBOTNot statedStock RMSE, MAPE and MSE
[26]LSTM10–50BOTS&P 500Stock MAPE
[41]SVM, RFNot statedVotingBSE SENSEXStock Accuracy
[42]NN2–5Not statedCIMB stock marketStock MSE, Accuracy
[44]SVM, LSTM and Multiple RegressionNot statedNot statedYahoo stock dataStock Accuracy
[45]NN30Not statedBrazilian stock marketStock Precision and recall
[29]NNNot statedBAGChinese stock marketStock Accuracy
[27]Tress and LSTM50–150BOT, STKS&P500 and NasdaqStock F-score, AUC and accuracy
[4]RNNNot statedSTK Stock AUC, accuracy
  1. CV cross-validation, RNN recurrent neural networks, CLF classification, MV majority voting, REG regression, MAPE mean absolute percentage error, MLP multi-layer perceptron, RMSE root mean square error