# Statistical Advisor, Searching for Patterns or Trends Over Time

## Time Series and Distributed Lags

Time Series and Distributed Lags chapter discusses full implementations of autocorrelation analysis, ARIMA, and Fourier analysis. In addition, a modeling 'work bench' is described which allows you to apply various transformations and smoothing techniques to the data. Such transformations often reveal trends or recurring patterns more clearly. The goal of ARIMA is to detect such seasonal patterns. Autocorrelation analysis allows you to examine the lagged effect (correlation) of a variable with itself or with another variable. Such effects may, for example, occur when correlating advertising expenditures over time with resultant sales; most likely, sales will be affected by increased advertising with some lag. The goal of Fourier analysis is to decompose a time series into simple underlying waves forms. All techniques make extensive use of graphics, allowing you to visualize any patterns.

Distributed Lags is a specialized option for analyzing the lagged effects of one independent (predictor, exogenous) variable with one dependent (criterion, endogenous) variable. Unlike the autocorrelation feature, this option can simultaneously evaluate multiple lags.

## Neural Networks

Neural Networks discusses numerous options for building neural networks to predict lagged response variables.