Amnesic Neural Network for Classification: Application on Stock Trend Prediction* Author: Qiang Ye, Bing Liang, Yijun Li				          Publication: ICSSSM 2005 				          Presenter:  Yu-Hsiang Huang2011.9.231
Introduction & Literature reviewMethodologyBP Neural Network Model Amnesic Neural Network Model Training AlgorithmExperiment Data Classification algorithm Experiment ResultOutline2
Artificial Neural Network models (ANN)Based on the neural structure of the brain
The ANN learns from experience
Critical step: network trainingTwo classes of predict stock market methodFundamental analysisMacroeconomic data and basic financial status of companyTechnical analysisHistory will repeat itselfThe correlation between price and volume reveals market behaviorPredictionBy exploiting implications hidden in past trading activities
By analyzing patterns and trends shown in price and volume chartIntroduction & Literature review3
Stock price predictionTraditional assumption: customer behavior is consistent
In real world: customer behavior change greatly
Training data set may be time-variant
Difficult to predict customer behavior from old dataTwo strategiesSelect all data from different time: can’t represent current knowledge
Select only the latest data: lose useful information hidden in data of early 			            timeData selectionin stock markets prediction will influence the training resultIntroduction & Literature review4
Amnesic Neural network (ANN*) modelIntroducing psychological notion of forgetting into Back Propagation (BP) neural networkSolve the problem of cross-temporal data classification
Effectiveness data depends on time
Present data is more useful than old data
Old data has less effect on training result, like gradually forgetting5Introduction (cont.)
Introduction & Literature reviewMethodologyBP Neural Network Model Amnesic Neural Network Model Training Algorithm Experiment Data Classification algorithmExperiment Result6
Artificial Neural NetworkComputational modeling tools for Modeling complex real-world problems
Capable of performing massively parallel computations for data processing and knowledge representation
The feed-forward-error-back-propagation learning algorithm is the most famous procedure for training ANNBack Propagation Neural NetworkBased on searching an error surface using gradient descent for points with minimum errors
Each iteration:
Forwardactivation to produce a solution
Backwardpropagation of computed error to modify the weight7Methodology
Introduction & Literature reviewMethodologyBP Neural Network Model Amnesic Neural Network Model Training Algorithm Experiment Data Classification algorithmExperiment Result8
9Methodology (cont.) BP network ModelInput:
Output:
Weight:      connects the node j in previous layer to the node k
Activation function:
Error signal:
I(n) is a set of input

Amnestic neural network for classification