Logged out
Access All our Essays in an Instant
Swap your work for FREE access, or pay £4.99 for instant access
Time Series Analysis and Forecasting
Member rating:
(2 votes)
| Words:
| Submitted: Thu Sep 30 2004
Page Preview
On the left is an image preview of every page of this document, and below are the first 150 words with formatting removed:
Time Series Analysis and Forecasting Coursework Plot 1. Time series plot of IBM price The time series plot shows an increasing trend, no seasonal or cyclical components. The autocorrelation function (acf) has high auto correlations that decrease slowly- giving the shape of a 'thick wedge'- this pattern is indicative of a trend. We cannot use our models because of the presence of the trend; this must be removed by differencing the series with a lag of 1. Plot 2. Time series plot of differenced IBM prices The plot and acf now show no pattern in the differenced series, we now have a stationary series on which to use our models. The first autocorrelation is significantly different from zero (T=2.43>2) and all the other autocorrelations lie within the confidence limits so this suggests that MA (1) model can be used for the differenced series. Only the first partial autocorrelation (pac) is significantly different from zero (T=2.43>2)...
Get instant access
- Instant, unlimited access to our documents in full
- Swap your work for free access, or pay £4.99
- To see the full version of this document and 147,383 others
Receive email updates for this category
- Simply tell us your email address and receive a weekly Study Help Email for FREE
- Receive 3 FREE essay views with each email
- Get all the latest essays from Coursework.Info & discussion from TheStudentRoom.co.uk
