On Ranking in Survival Analysis: Bounds on the Concordance Index
نویسندگان
چکیده
In this paper, we show that classical survival analysis involving censored data can naturally be cast as a ranking problem. The concordance index (CI), which quantifies the quality of rankings, is the standard performance measure for model assessment in survival analysis. In contrast, the standard approach to learning the popular proportional hazard (PH) model is based on Cox’s partial likelihood. We devise two bounds on CI–one of which emerges directly from the properties of PH models–and optimize them directly. Our experimental results suggest that all three methods perform about equally well, with our new approach giving slightly better results. We also explain why a method designed to maximize the Cox’s partial likelihood also ends up (approximately) maximizing the CI.
منابع مشابه
Support vector methods for survival analysis: a comparison between ranking and regression approaches
OBJECTIVE To compare and evaluate ranking, regression and combined machine learning approaches for the analysis of survival data. METHODS The literature describes two approaches based on support vector machines to deal with censored observations. In the first approach the key idea is to rephrase the task as a ranking problem via the concordance index, a problem which can be solved efficiently...
متن کاملEfficiency Evaluation and Ranking DMUs in the Presence of Interval Data with Stochastic Bounds
On account of the existence of uncertainty, DEA occasionally faces the situation of imprecise data, especially when a set of DMUs include missing data, ordinal data, interval data, stochastic data, or fuzzy data. Therefore, how to evaluate the efficiency of a set of DMUs in interval environments is a problem worth studying. In this paper, we discussed the new method for evaluation and ranking i...
متن کاملRanking of decision making units based on cross efficiency by undesirable outputs and uncertainity
Cross efficiency is one of the useful methods for ranking of decision making units (DMUs) in data envelopment analysis (DEA). Since the optimal solutions of inputs and outputs weights are not unique so the selection of them are not simple and the ranks of DMUs can be changed by the difference weights. Thus, in this paper, we introduce a method for ranking of DMUs which does not have a unique pr...
متن کاملInvestigating Effect of Oil Revenues on Social Capital in Oil Rentier Countries: An Extreme Bounds Analysis Approach (EBA)
The volatility of oil revenues in rentier countries, due to the large share of oil revenue in their economies, affects many variables. Oil revenues can also affect social behavior and culture. Therefore, social capital, as the result of social networks and institutions, can be directly and indirectly affected by oil rents. The review of economic literature suggests that oil revenues could poten...
متن کاملBounds on First Reformulated Zagreb Index of Graph
The first reformulated Zagreb index $EM_1(G)$ of a simple graph $G$ is defined as the sum of the terms $(d_u+d_v-2)^2$ over all edges $uv$ of $G .$ In this paper, the various upper and lower bounds for the first reformulated Zagreb index of a connected graph interms of other topological indices are obtained.
متن کامل