E.g. 9.1 Predicting the interval to eruption of Old Faithful Geyser A geyser is a hot spring that occasionally becomes unstable and erupts hot water and steam into the air. Different geysers will erupt for various lengths of time and intervals. Probably the most famous geyser is Old Faithful, in Yellowstone National Park, Wyoming. The intervals between eruptions of Old Faithful range from about 30 to 90 minutes. Water shoots to heights generally over 35 meters, with eruptions lasting from 1 to 5.5 minutes (Marler, 1969). Because of its regularity and beauty, Old Faithful Geyser is a major tourist attraction, and it is not uncommon for several thousand people to watch an eruption on a summer afternoon. Prediction of the time of the next eruption of this geyser is of interest both to the Park Service and to visitors. In fact, prediction of the time of the next eruption is posted in prominent locations near the geyser. Given opposite are i = interval to the next eruption and dur = duration of an eruption, both in minutes, for all eruptions of Old Faithful Geyser between 6 a.m. and midnight for August 1 (day=1) to August 8 (day=8), 1978 Data on Old Faithful has been collected for many years by ranger naturalists in the park, using a stopwatch. The duration measurements have been rounded to the nearest 0.1 minutes or 6 seconds, while intervals are reported to the nearest minute. The National Park Service uses values of x to predict future values for y, and we shall follow their example. It is unlikely that a causal relationship exists between x and y, rather both are probably results of some other unobserved cause or set of causes. Modeling the observed association between x and y may lead to useful predictions, but the fitted relationship may have no geological meaning. Source: Weisberg, S. Applied Linear Regression, 2nd Edn. Wiley, New York, 1985. Excerpts... day interval duration year(1978=0, 1979=1) 1 78 4.4 0 1 74 3.9 0 1 68 4.0 0 .. .. 8 77 4.2 0 8 73 4.4 0 .. 8 80 4.8 0 8 79 4.1 0 Data for Validation: 119 eruptions of Old Faithful for the period August 16 to August 23, 1979, a year after the original data.( in same format and same dataset as above) 16 82 4.1 1 16 80 4.2 1 16 76 4.5 1 .. 23 81 4.2 1 23 48 2.1 1 23 84 5.2 1 23 63 2.0 1