cA39QuiIrQg Bayesian - Credit - Data - Analysis and Chi Quared
cA39QuiIrQg Bayesian - Credit - Data - Analysis and Chi Quared
Probabilty of Default 5%
Gender H0 Defaults are distributed evenly
Expected nDefaults Defaults Totals Ha Defaults are not distributed evenly
Male 665.00 35.00 700.00
Female 285.00 15.00 300.00
950.00 50.00 1000.00
Chi-squared 1.60
Treshold 95%
# DF 1
Cut-off point 3.84
We accept the 0-Hypothesis and conclude that gender does not have a predictive value
Activity
Expected nDefaults Defaults Totals
Agriculture 172.90 9.10 182.00
Commerce 164.35 8.65 173.00
Industry 83.60 4.40 88.00
Services 255.55 13.45 269.00
Crafts 273.60 14.40 288.00
950.00 50.00 1,000.00
Chi-squared 48.78
Treshold 95%
# DF 4
Cut-off point 9.49
We reject the 0-Hypothesis and conclude that "Activity" does have a predictive value
Age
Expected nDefaults Defaults Totals
Under twenty 162.45 8.55 171
Twenties 228 12 240
Thirties 238.45 12.55 251
Fourties 181.45 9.55 191
Over fifty 139.65 7.35 147
950 50 1000
Chi-squared nDefaults Defaults
Under twenty 3.68 69.92
Twenties 0.28 5.33
Thirties 0.31 5.82
Fourties 0.17 3.23
Over fifty 0.04 0.75
Chi-squared 89.53
Treshold 95%
# DF 4
Cut-off point 9.49
We reject the 0-Hypothesis and conclude that "Activity" does have a predictive value
Family situation
Expected nDefaults Defaults Totals
Single 139.65 7.35 147
Married w/o children 115.90 6.10 122
Married with children 694.45 36.55 731
Totals 950 50 1000
Chi-squared 43.85
Treshold 95%
# DF 2
Cut-off point 5.99
We reject the 0-Hypothesis and conclude that "Family situation" does have a predictive value
distributed evenly
not distributed evenly