Stocks Selection Using Neural Network

1
School of Electrical and Electronic
Engineering Information Communication Institute
of Singapore
Stocks Selection Using Neural Network
Objectives
Approach

• To utilize Neural Network to determine the
  factors that affect the pricing of stocks.
• To study the use of Neural Network in selecting
  and ranking stocks for equity investment.

• Analyze and study the financial,
  sectorial/industrial, local, regional and
  worldwide macro-economic factors that potentially
  affect equity markets.
• Select companies for the experiment from every
  sector based on pre-defined criteria such as
  market capitalization.
• Perform data preparation and correlation analysis
  of inputs gathered.
• Train the GRNN using selected inputs.
• Analyze the neural networks ability to predict
  the performance of stocks prices.
• Use GRNN to select stocks that will yield good
  returns for investors.

Introduction

• Stock pricing has always been a result of
  multiple factors such as company financials,
  macro-economic factors and rivals performance.
• Researchers had tried different techniques in
  predicting the up and down of stocks prices.
  Among common technique used are technical
  analysis, fundamental analysis and artificial
  intelligent assisted techniques.
• Generalized Regression Neural Network (GRNN) is a
  subclass of radial basis networks and has proven
  its usefulness in many applications domains such
  as forecasting and function estimation.

Expected Results

• A set of factors that will have significant
  contributions to stock price movements.
• A trained neural network that will be able to
  assist in selecting stocks for investment.

TEACH / USE
Y1
X1
W1
Data pre-processing
W2
Y2
X2
Neural Network
Selected stocks
Xn
Ym
Wn
mgtn
Student Wong Ka Yew
Supervisor Dr. Quah T.S, Jon
Information Communication Institute of Singapore
? Blk S2, Level B3c-13, Nanyang Avenue,
Singapore 639798 ? Website http//www.icis.ntu.
edu.sg