Evaluation to determine which of the following three models yields the best predictive result when analyzing historical stock data.
-
Neural Network
-
A LSTM RNN model to predict entry and exit points that might generate profitable trades
-
Sequential model with four layers
-
-
Time-Series
-
Univariate time series modelling using ARIMA to forecast closing stock price.
-
Multivariate time series modelling with correlated assets and sentiment scores as dependent variables using ARIMA.
-
-
Decision Tree and Random Sampling
-
Create a decision tree model to determine entry & exit point of the selected public equity
-
Determine the precision through multiple random sampling model
-
The key takeaway is that it is extremely difficult to predict stock returns.