张俊妮《概率统计》北大光华-lecture18.pptVIP

Example Consider a particular automobile model that currently attains an average fuel efficiency of 10 liters per 100 kilometers. A product research group developed a new fuel injection system (燃油喷射系统) specifically designed to decrease the liters-per-100-kilometers rating. To evaluate the new system, 30 were manufactured, installed in automobiles, and subjected to research-controlled driving tests. What is your conclusion? What is your conclusion? Statistical Inference Type 3: Hypothesis Testing (假设检验) Which decision? Which decision? Concept: Level of Significance (显著性水平) The level of significance is the maximum probability of type I error that is allowed, and is denoted by a0. The p-value (p值) Approach to Hypothesis Testing A p-value is a probability that provides a measure of the evidence against the null hypothesis provided by the sample. Smaller p-values indicate more evidence against . Benefit of Using the p-value Approach p value can be compared with multiple values of a0, and we therefore do not need to specify level of significance at the beginning. E.g. p value=0.027, what is the decision when a0 =0.01, a0 =0.05? Hence, p value is standard output in statistical softwares. * An Example of Hypothesis Testing Lecture 18 9.71 8.81 10.20 9.69 11.00 9.26 10.20 9.61 9.58 9.13 9.34 10.70 7.83 9.95 11.00 9.93 9.93 9.30 10.40 9.07 10.00 9.44 10.70 11.50 9.17 6.91 10.10 10.70 8.20 9.50 Suppose that the results from driving tests of the 30 automobiles are as follows: Suppose that the results from driving tests of the 30 automobiles are as follows: 9.71 8.81 10.20 9.69 11.00 9.26 10.20 9.61 9.58 9.13 9.34 10.70 7.83 9.95 11.00 9.93 9.93 9.30 10.40 9.07 10.00 9.44 10.70 11.50 9.17 6.91 10.10 10.70 8.20 9.50 30 30 or 30 30 30 30 30 30 30 30 or What does it mean? What’s the consequence? What does it mean? What’s the consequence? Type I Error (第一类错误) and Type II Error (第二类错误) Two Types of Errors Probability of making Type I error Probabili