Abstract: The uniequness theorem for the Tsallis entropy by introducing the generalized Faddeev's axiom is proven. Our result improves the recent result, the uniqueness theorem for Tsallis entropy by the generalized Shannon-Khinchin's axiom in \cite{Suy}, in the sence that our axiom is simpler than his one, as similar that Faddeev's axiom is simpler than Shannon-Khinchin's one.
What problem does this paper attempt to address?
### Problems the paper attempts to solve
This paper aims to solve the problem of estimating the causal effects of time - dependent treatments on survival time or other continuously distributed outcomes in continuous time. Specifically, the paper focuses on how to estimate the treatment effects when the treatment is repeatedly adjusted according to the patient's time - dependent characteristics.
### Background and challenges
1. **Time - dependent confounding factors**:
- The treatment effect cannot be simply estimated by conditioning on these time - dependent characteristics, because these characteristics themselves may be indicators of the treatment effect. Such time - dependent confounding factors are very common in observational studies.
- For example, doctors may adjust the treatment plan according to the patient's health condition, which makes simple regression analysis unable to accurately estimate the treatment effect.
2. **Limitations of existing methods**:
- Existing methods such as linear regression, logistic regression or Cox regression have limitations in dealing with time - dependent confounding factors.
- Even the time - dependent Cox model cannot directly answer whether the treatment affects the outcome of interest, because past covariate values may be affected by previous treatments.
### Solutions
1. **Structural nested models**:
- Robins proposed structural nested models (SNMs) to estimate treatment effects in the presence of time - dependent confounding factors.
- These models assume that all the information used by doctors to make treatment decisions and the information for predicting the final outcome of patients are analyzable. This assumption is called "no unmeasured confounding factors".
2. **Extension in continuous time**:
- This paper provides the conceptual framework and mathematical formalization of structural nested models in continuous time.
- It is proved that the estimators obtained under this framework are consistent and asymptotically normal.
3. **Testing methods**:
- A method is proposed to test whether the treatment affects the outcome of interest without specifying the treatment - effect model.
### Main contributions
1. **Theoretical basis**:
- It provides a solid theoretical basis for the application of structural nested models in continuous time.
- It proves the consistency and asymptotic normality of the estimators of structural nested models in continuous time.
2. **Practical applications**:
- It illustrates the application of these methods through a specific example.
- It provides an effective method for dealing with time - dependent confounding factors, especially in observational studies in medical research.
### Conclusion
By introducing the framework of structural nested models in continuous time, this paper solves the difficult problem of estimating treatment effects in the presence of time - dependent confounding factors, and provides a new tool and method for observational studies in medical research.