1: \begin{abstract}

2: We propose a new approach for estimating the parameters of a probability

3: distribution. It consists on combining two new methods of estimation. The

4: first is based on the definition of a new distance measuring the difference

5: between variations of two distributions on a finite number of points from

6: their support and on using this measure for estimation purposes by the method

7: of minimum distance. For the second method, given an empirical discrete

8: distribution, we build up an auxiliary discrete theoretical distribution

9: having the same support of the first and depending on the same parameters of

10: the parent distribution of the data from which the empirical distribution

11: emanated. We estimate then the parameters from the empirical distribution by

12: the usual statistical methods. In practice, we propose to compute the two

13: estimations, the second based on maximum likelihood principle of known

14: theoretical properties, and the first being as a control of the effectiveness

15: of the obtained estimation, and for which we prove the convergence in

16: probability, so we have also a criterion on the quality of the information

17: contained in the observations. We apply the approach to truncated or grouped

18: and censored data situations to give the flavour on the effectiveness of the

19: approach. We give also some interesting perspectives of the approach including

20: model selection from truncated data, estimation of the initial trial value in

21: the celebrate \textit{EM} algorithm in the case of truncation and merged

22: normal populations, a test of goodness of fit based on the new distance,

23: quality of estimations and data.

24: 

25: \end{abstract}

26: