कैपलन-मेयर द्वारा सर्वाइवल कर्व विश्लेषण

R में Marketing Analytics के लिए Machine Learning

Verena Pflieger

Data Scientist at INWT Statistics

सर्वाइवल ऑब्जेक्ट I

cbind(dataSurv %>% select(tenure, churn),
      surv = Surv(dataSurv$tenure, dataSurv$churn)) %>% head(10)

   tenure churn surv
1       1     0   1+
2      34     0  34+
3       2     1   2 
4      45     0  45+
5       2     1   2 
6       8     1   8 
7      22     0  22+
8      10     0  10+
9      28     1  28 
10     16     0  16+

कैपलन-मेयर विश्लेषण

fitKM <- survfit(Surv(dataSurv$tenure, dataSurv$churn) ~ 1, 
                 type = "kaplan-meier")
fitKM$surv

 [1] 0.9284504 0.9045343 0.8859371 0.8692175 0.8561374
 [6] 0.8478775 0.8372294 0.8283385 0.8184671 0.8086794
[11] 0.8018542 0.7933760 0.7847721 0.7792746 0.7707060
[16] 0.7641548 0.7580075 0.7522632 0.7476436 0.7432153
[21] 0.7389925 0.7321989 0.7288777 0.7228883 0.7168003
[26] 0.7127809 0.7092320 0.7059049 0.7016930 ...

survfit ऑब्जेक्ट प्रिंट करना

print(fitKM)

Call: survfit(formula = Surv(dataSurv$tenure, dataSurv$churn) ~ 1, 
    type = "kaplan-meier")

      n  events  median 0.95LCL 0.95UCL 
   5311    1869      70      68      72

plot(fitKM)

श्रेणीगत कोवेरिएट के साथ कैपलन-मेयर

fitKMstr <- survfit(Surv(tenure, churn) ~ Partner, 
                    data = dataSurv)

print(fitKMstr)

Call: survfit(formula = Surv(tenure, churn) ~ Partner,
              data = dataSurv)

               n events median 0.95LCL 0.95UCL
Partner=No  2828   1200     45      41      50
Partner=Yes 2483    669     NA      NA      NA

plot(fitKMstr, lty = 2:3)
legend(10, .5, c("No", "Yes"), lty = 2:3)

अभ्यास करते हैं!

R में Marketing Analytics के लिए Machine Learning