Outliers or extreme values are patterns in the data, which do not conform to a well-defined concept of normal behavior. In today's often changing environment, detecting and forecasting outliers in time series related to stock market, credit card fraud, fraud in insurance systems, tourism demand indicators, etc., is a challenge for both humans and computers. In this paper, we present, for the first time, the association among the outliers in different univariate time series, and we formally define Mining Association of Extreme Values (MAEV). We then investigate how MAEV can be applied to forecasting outliers in a time series based on the detection of outliers in another. We evaluate the efficiency of the proposed methodology by applying it to hotel booking demand. More specifically, we first use an algorithm for automatically detecting outliers in time series such as booking volumes, arrival volumes, or booking cancellations, then we form a set of instances that correspond to time intervals, by considering in each instance the existence or not of an outliers for every different time series, then we apply Apriori association rule mining algorithm to the formed set of instances, and finally, we use the extracted association rules to forecast more outliers.