张俊妮《概率统计》北大光华-lecture14.pptVIP

Visual Inspection: Histogram (直方图) Key Findings Actions Key Findings: Mean is a simple measure for “location”. It is very sensitive to outliers. There might be other way to measure “location”. Variability is the missing key! Actions: Are there other ways to measure location? How to measure variability? Scatter Diagram (散点图) Qualitative Variables (定性变量) Bar Chart for the “Education” Variable Pie Chart for the “Education” Variable Examples of More Advanced Problems What is the profile of all peasant workers in Beijing? Distribution of education Mean number of household members Proportion of having social insurance Proportion of having employment contract Mean monthly household income Mean monthly household consumption … Examples of More Advanced Problems How does monthly household income of peasant workers in Beijing depend on Education Number of Household members Whether having social insurance Whether having employment contract Examples of More Advanced Problems Prediction! A bank: using internal models to predict probability of default. Historical data about companies getting loans from the bank dfghdfghfdghgfhdfghdfgh * dfghdfghfdghgfhdfghdfgh * dfghdfghfdghgfhdfghdfgh * * Quantitative Variable: observations measured on a naturally occurring numerical scale. Qualitative Variable: observations can only be classified into one of a group of categories. Qualitative Variables in the Example: Education Insurance Contract E.g. The “Education” Variable 2.5 0.025 36 5 10.2 0.102 144 1 45.7 0.457 647 2 31.4 0.314 445 3 Education frequency relative frequency percent frequency 0 33 0.023 2.3 4 111 0.078 7.8 Total 1416 1 100 Total Insurance 36 23 13 5 647 152 495 2 445 121 324 3 111 53 58 4 Education 0 1 0 27 6 33 1 119 25 144 Total 1036 380 1416 Cross-tabulation of “Education” and “Insurance” Total Insurance 100 63.9 36.1 5 100 23.5 76.5 2 100 27.2 72.8 3 100 47.7 52.3 4 Education 0 1 0 81.8 18.2 100 1 82.6 17.4 100 Cross-tabulation of “Education” and “Insu