Your SlideShare is downloading.
×

×

Saving this for later?
Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.

Text the download link to your phone

Standard text messaging rates apply

Like this presentation? Why not share!

- Neural network & its applications by Ahmed_hashmi 54042 views
- Artificial intelligence NEURAL NETW... by REHMAT ULLAH 10882 views
- Stock Ranking - A Neural Networks A... by riturajvasant 4063 views
- pattern recognition in stock market by Zorro29 4294 views
- Automatic time series forecasting by Rob Hyndman 2746 views
- Introduction Of Artificial neural n... by Nagarajan 6409 views

3,219

Published on

The aim of the project is to predict the interest rates,bond yield variation and stock market prices using neural networks and make a comparative study of different pre-processing techniques viz Fast …

The aim of the project is to predict the interest rates,bond yield variation and stock market prices using neural networks and make a comparative study of different pre-processing techniques viz Fast Fourier Transform and Hilbert Huang Transform.

this ppt needs other two also..

Published in:
Technology

No Downloads

Total Views

3,219

On Slideshare

0

From Embeds

0

Number of Embeds

0

Shares

0

Downloads

171

Comments

0

Likes

2

No embeds

No notes for slide

- FINANCIAL FORECASTING USING NEURAL NETWORKS <br />Presented by , Amit jain 07000519Ranjeet ranjan 07000537puneet gupta 07000534<br />
- What is Financial Forecasting<br />Prediction of prices of instruments of speculation<br />Stocks<br />Commodity futures<br />Exchange Rates<br />Interest Rates .<br />Problem : Non linear and non stationary data<br />
- Methods Used<br />Fundamental Analysis<br />Understanding the supply demand curve<br />Involves studying of news and economic factors <br />Technical Analysis<br />Understanding historical price patterns<br />Tools like moving average, learning systems<br />Latest Approach: Combine Technical and Fundamental Analysis<br />
- NEURAL NETWORKS <br />Map some type of input stream of information to an output stream of data. <br />They derive non-linear modelsthat can be trained to map past and future values of the input output relationship .It extracts relationships governing the data that was not obvious using other analytical tools. <br />Capability to recognize patternand the speed of techniques to accurately solve complex processes, exploited exhaustively in financial forecasting.<br />Trained without the restriction of a model to derive parameters and discover relationships, driven and shaped solely by the nature of the data. <br />
- NEURAL NETWORKS V/S CONVENTIONAL COMPUTERS<br />Neural networks have the unique capability of learning thus are adaptive .This problem solving tool creates a unique likeness to the human brain .<br />Use the interconnectedness of the elements of the model rather than follow a set of sequential steps, that may or may not solve the problem like computers do.<br />A different aspect of model building, where the unique relationships between the variables creates the model, rather than trying to force variables to conform to a theoretical abstract that may or may not exist.<br />
- NEURAL NETWORKS IN FINANCE<br />Neural networks are trained without the restriction of a model to derive parameters and discover relationships, driven and shaped solely by the nature of the data. Thus it has profound implications and applicability to the finance field.<br />Some of the fields where it is applied are: <br /> Financial forecasting<br /> Capital budgeting and risk management <br /> Stock market analysis<br />Used to analyze and verify Economic hypothesis and theories which were not possible otherwise. <br />Govt. predicts interest rates to gauge the future inflationary situation of its economy .<br />
- Neural Networking and Similarities with the Workings of the Human Brain<br />
- A SIMPLE NEURON<br />
- VECTOR INPUT TO NEURON<br />
- LAYER OF NEURONS <br />
- LAYER OF NEURONS …..<br />
- MULTIPLE LAYERS <br />
- MULTIPLE LAYERS …..<br />
- NARX MODEL<br />
- TRANSFER FUNCTIONS <br />
- TRAINING ALGORITHMS<br />trainlm : fastest and better for non-linear cases , default for feed-forwardnet .<br />
- BACK-PROPOGATION<br />Numerous such input/target pairs are used to train the Neural Network.<br />
- TIME SERIES FORECASTING<br />Time series forecastingor time series prediction, takes an existing series of data and forecasts the data values. The goal is to observe or model the existing data series to enable future unknown data values to be forecasted accurately.<br />Done using the NARX model or NAR model .<br />
- DIFFICULTIES<br />Limited quantity of data .<br />Noise in data – It obscures the underlying pattern of the data .<br />Non-stationarity - data that do not have the same statistical properties (e.g., mean and variance) at each point in time<br />Appropriate Forecasting Technique Selection .<br />
- Preprocessing of Training Data<br />Reason: Need to understand underlying patterns.<br />Tools:<br />Moving Average<br />Fast Fourier Transform (FFT)<br />Hilbert Huang Transform (HHT)<br />
- Types Of Data Worked Upon<br />Interest Rates (RBI 91 day Govt. Of India Treasury Bills)<br />Sensex Data ( 2005-2010)<br />Exchange Rates (Daily Exchange Rates of INR-Dollars 2004-2011)<br />All the Data are divided into Three Sets<br />Training Set<br />Testing Set<br />Validation Set<br />
- Types Of Preprocessing <br />No Pre-Processing (Simple NN)<br />Using FFT (FFT NN)<br />Using HHT (HHT NN)<br />All the types of data are used on all the types of preprocessing techniques , therefore generating 9 cases.<br />Now, we Compare all of them Data-Wise.<br />
- 1. Interest Rates<br />The interest rate data is applied on all three kinds of preprocessing. The Error Graphs are as:<br />Simple NN<br />
- FFT NN<br />
- HHT NN<br />
- 2. Sensex Data<br />The sensexdata is applied on all three kinds of preprocessing. The Error Graphs are as:<br />Simple NN<br />
- FFT NN<br />
- HHT NN<br />
- 3. Exchange Rates<br />The Exchange Rate data is applied on all three kinds of preprocessing. The Error Graphs are as:<br />Simple NN<br />
- FFT NN<br />
- HHT NN<br />
- Conclusion from Results<br />Pre-processing can boost the Neural Network Performance<br />The performance of Neural Network also depends on the nature of the data series<br />
- Thank You<br />

Be the first to comment