A Fully Integrated Environment for Time-Dependent Data Analysis
Time Series performs univariate and multivariate analysis
and enables you to explore both stationary and nonstationary models.
You can select a model to fit your data and obtain estimates of the
model's parameters. Choose from standard methods such as Yule-Walker,
Levinson-Durbin, long autoregression, Hannan-Rissanen, and others.
After reading in and plotting your data, use the built-in Time
Series transforms for linear filtering, simple exponential
smoothing, differencing, moving averages, and more to transform your
raw data into a form suitable for modeling. Calculating and plotting
the correlation and partial correlation functions will help you spot
patterns. Once you select a model to fit your data, Time Series
makes it easy to estimate the model parameters and check its validity
using residuals and tests such as the portmanteau, turning points,
difference-sign, and others.
If you need to predict future values, Time Series can help
there, too. Best linear predictor and approximate best linear predictor
are among the commonly used forecasting techniques included. Collect
new data and you can instantly update your predictions.
In addition, Time Series enables you to analyze your data in
frequency space. The spectral analysis tools inside Time Series
use the Fourier transform and other robust numerical methods.
This package is also an ideal instructional tool with its
description of the fundamentals of time series analysis and its clear,
concise examples.
The package comes with electronic documentation, which is fully
integrated with the Wolfram Mathematica Documentation Center.
Time Series 1.4
requires Mathematica 7 and is available for all Mathematica
platforms.
|