Shulou(Shulou.com)06/01 Report--

How to carry out survival analysis of R language in big data, for this problem, this article introduces the corresponding analysis and solution in detail, hoping to help more small partners who want to solve this problem find a simpler and easier way.

I. Basic concepts

Survival analysis: To study whether each factor has relationship with survival time and the degree of association. Can be extended to disease recurrence time, machine failure time, etc.

Initiating event: An initial characteristic event that reflects the beginning of the survival process of the subject.

Endpoint event (death event): occurrence of a specific outcome of interest to the investigator. For example,"the patient died of the disease."

Observation time: The time from the start of observation to the end of observation.

Survival time: Observed survival time, denoted by the symbol t.

Complete data: Time elapsed from observation to death event, survival time is complete.

Censored data (censored value): Observation time does not end due to endpoint event, but ends due to (1) lost to follow-up (2) death due to non-study factors (3) observation end while subject is still alive. Often a "+" is placed in the upper right corner of truncated data to indicate that the object is likely to live longer.

Variables: There are two variables for survival analysis: survival time t and outcome variable (0-1). Where outcome variable 1 indicates a death event and 0 indicates censoring.

Survival curve: Draw a survival curve with survival time as horizontal axis and survival rate as vertical axis.

Median survival time: Also called median survival time, it means that exactly 50% of individuals survive longer than this time.

Part of the Survival Analysis Study

1. Describe the survival process

To study the distribution characteristics of survival time, estimate survival rate and its standard error, and draw survival curve. Common methods include product limit method (K-M method) and life table method. Here, product limit method (K-M method) is briefly introduced.

http://www.360doc.com/content/17/0626/11/6175644_66623573.shtml Here is a graphic illustration of what KM curves can tell us and the fundamentals of the product limit method.

Note: Survival curves are univariate analyses (two or more levels), with median survival times representing mean survival times;

2. Compare the survival process

After obtaining estimates of survival rates and their standard errors, comparisons of survival curves between two or more groups can be made, often using log-rank tests. If the curves cross, there may be confounding factors.

Note: Comparison of two or more survival curves (two or more levels for a single factor)

Analysis of factors affecting survival time

Cox proportional hazards regression model is a common method for multivariate survival analysis. Maybe it'll be a long story.

III. R package for survival analysis

1. R-survival package for survival analysis, and draw KM curve:

Surv: Used to create survival data objects

survfit: Create KM survival curves or Cox adjusted survival curves

survdiff: statistical test for different groups

coxph: Construct COX regression model (to be)

cox.zph: Test PH hypothesis (pending)

#Load package, read data

library(survival)

dat

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.