Google just published a very interesting article "On the predictability of Search Trends", where past queries are used for predicting the trends of future queries. I like this paper.
"Specifically, we have used a simple forecasting model that learns basic seasonality and general trend. For each trends sequence of interest, we take a point in time, t, which is about a year back, compute a one year forecasting for t based on historical data available at time t, and compare it to the actual trends sequence that occurs since time t. The error between the forecasting trends and the actual trends characterizes the predictability level of a sequence, and when the error is smaller than a pre-defined threshold, we denote the trends query as predictable."
This is a more approach that you can use in many contexts. For instance, I have seen it used for understand the coverage and the precision of firing a vertical result time based (such as news, blogs, twitter) into the SERP.
Another observation about the paper. You can better predict wether a query has a predictable trend, by enriching your Querylog with other temporal based data such as Twitter, News and blogs.