Analyzing Time-Series Data: What Model to Use?

[colby2152]

I want to test a series of 36 time periods and 10 data points for each one of those periods. The goal is to see which data points have the biggest effect on change from one period to the next.

I thought Chi-Square was the way to go, but I do not have any expected frequencies here. Regression analysis is more of a prediction/forecasting model? What statistical model should I use here?

 

[EnumaElish]

A decent stats package will provide outlier detection as part of their basic regression subroutine.

For SAS, see:http://www2.sas.com/proceedings/sugi24/Infovis/p161-24.pdf and http://www2.tltc.ttu.edu/westfall/images/5349/outliers_what_to_do.htm

Your regression model could be y(i, t) = a + b y(i, t-1) where y(i,t) (i = 1, ..., 10) corresponds to the 10 different observations in period t. (Similarly for t-1.)

 

[tacman]

There are a few different statistical models that could potentially be used to analyze this type of time-series data. The choice of model will depend on the specific goals and questions of your analysis, as well as the characteristics of your data.

One possible approach could be to use a regression model, as you mentioned. This would allow you to examine the relationship between your independent variable (data points) and your dependent variable (change from one period to the next). You could also use regression to make predictions about future changes based on your data.

Another option could be to use a time-series analysis model, such as ARIMA (Autoregressive Integrated Moving Average). This type of model is specifically designed for analyzing trends and patterns in time-series data. It takes into account the autocorrelation and seasonality of the data, which may be important in understanding the changes between periods.

Alternatively, you could also consider using a multivariate analysis approach, such as multivariate regression or multivariate time-series analysis. These models allow you to examine the relationship between multiple independent variables and your dependent variable, which could be useful if you have more than one data point that you believe may be influencing the change between periods.

Ultimately, the best model to use will depend on the specific goals and questions of your analysis, as well as the characteristics of your data. It may be helpful to consult with a statistician or data analyst to determine the most appropriate model for your particular research question.