The construction of generalized B-spline low-pass filters related to possibility density (Q1934380): Difference between revisions

Let \(\lambda \in (0,\,1)\) and let \((c_k)_{k=0}^{\infty}\) be a real sequence with \(|c_k| \leq c\, \lambda^k\) \((k=0,1,\ldots)\). The authors introduce the so-called generalized symmetric B-spline of order \(n\in \mathbb N\) by the recursive convolution formula \(\beta_n = \beta_{n-1}\ast \beta_1\) \((n=2,3,\ldots)\), where \[ \beta_1 (t) = \sum_{k=0}^{\infty} c_k \,\chi_{(-(k+1)/2,-k/2]\cup [k/2,(k+1)/2)}(t) \quad (t\in \mathbb R) \] and \(\chi_M\) is the indicator function of the set \(M\). In the case \(c_0=1\), \(c_k=0\) \((k=1,2,\ldots)\) one obtains the classical symmetric B-spline. It is shown that the properties of \(\beta_n\) are generalizations of some known properties of the classical symmetric B-spline.