In this appendix, we will develop in detail the estimation process of the model parameters (rules probabilities,
\(\,p(A \rightarrow \alpha ) \in {\mathcal {P}} \;\text {of}\; {\mathbb {G}},\,\) see Def. (
3)) until reaching expression (
9). To begin with, we present the growth transformation optimization framework and how to optimize Eq. (
8) by applying the growth transformations for rational functions [
16]. In each iteration of the optimization process, the update parameters
\(\,p(A \rightarrow \alpha )\,\) are obtained using the following expression:
$$\begin{aligned} {\bar{p}}(A \rightarrow \alpha )\,=\,\frac{p(A \rightarrow \alpha )\, \left[ \frac{\partial \, {Q}_{\pi }({\mathbb {G}}, \Omega )}{\partial \, p(A \rightarrow \alpha )} + \textrm{C}\right] _{\pi }}{\sum _{i=1}^{n_{A}}\; p(A \rightarrow \alpha _{i}) \left[ \frac{\partial \, {Q}_{\pi }({\mathbb {G}}, \Omega )}{\partial \, p(A \rightarrow \alpha _{i})} + \textrm{C}\right] _{\pi }} \end{aligned}$$
(12)
\(n_A\) is the number of rules with the non-terminal
A as the left side of the rule, and
\(\,\pi = (\pi _{A_1}, \pi _{A_2}, \ldots , \pi _{A_{\vert N\vert }})\):
\(\,A_i \in N\),
\(\,1 \le i \le \vert N\vert \;\) is a vector defined as follows:
\(\,\pi _{A_i} = (p(A_i \rightarrow \alpha _{i1}), p(A_i \rightarrow \alpha _{i2}), \ldots , p(A_i \rightarrow \alpha _{in_{A_i}}))\). Furthermore,
\({Q}_{\pi }({\mathbb {G}}, \Omega )\) (see Eq. (
8)) is a polynomial function. As it was demonstrated in [
16], for every point of the domain
\(\pi\), there is a constant
\(\,\textrm{C}\,\) such that the polynomial
\(\,P_{\pi }+\textrm{C}\,\) has only non-negative coefficients. Following a similar development to that used in [
11], we will allow us to obtain
\(\,{\bar{p}}(A \rightarrow \alpha )\,\) from Eq. (
12).
First of all, let us define an auxiliary function
$$\begin{aligned} {\mathcal {D}}_{A \rightarrow \alpha }^h (\Delta _x)\,=\, p(A \rightarrow \alpha )\, \left[ \,\frac{\partial \,\prod _{x\in \Omega } P (x, \Delta _x)^h}{\partial \, p(A \rightarrow \alpha )}\,\right] _{\pi } \end{aligned}$$
then expression (
12) can be rewritten as
$$\begin{aligned}&{\bar{p}}(A \rightarrow \alpha )\,=\, \nonumber \\&\;\frac{{\mathcal {D}}_{A \rightarrow \alpha }^1\,(\Delta _x^r)\, -\ \left( {\widetilde{F}}_h({\mathbb {G}}, \Omega )\right) _{\pi }\;\; {\mathcal {D}}_{A \rightarrow \alpha }^h\,(\Delta _x^c)\, +\ p(A \rightarrow \alpha )\;\; \textrm{C}}{\sum _{i=1}^{n_{A}}\;{\mathcal {D}}_{A \rightarrow \alpha _i}^1\,(\Delta _x^r) -\left( {\widetilde{F}}_h({\mathbb {G}}, \Omega )\right) _{\pi } \sum _{i=1}^{n_{A}}\, {\mathcal {D}}_{A \rightarrow \alpha _i}^h\,(\Delta _x^c) + p(A \rightarrow \alpha _i)\,\textrm{C}} \end{aligned}$$
(13)
We will begin by developing the expression
\(\,{\mathcal {D}}_{A \rightarrow \alpha }^h (\Delta _x)\,\) as a preliminary step to evaluating the expressions
\(\,{\mathcal {D}}_{A \rightarrow \alpha }^1 (\Delta _x^r)\,\) and
\(\,{\mathcal {D}}_{A \rightarrow \alpha }^h (\Delta _x^c)\,\) of the numerator.
$$\begin{aligned}&{\mathcal {D}}_{A \rightarrow \alpha }^{h} (\Delta _{x})\,=\nonumber \\&\quad =\, p(A \rightarrow \alpha )\,\left[ \,h\;\prod _{x \in \Omega } {P} (x, \Delta _{x})^{h-1}\; \frac{\partial \prod _{x \in \Omega }\,{P} (x, \Delta _{x})}{\partial \, p(A \rightarrow \alpha )}\,\right] _{\pi } \nonumber \\&\quad =\,h\,\left[ \,\prod _{x \in \Omega }\, {P} (x, \Delta _{x})^{h}\;\sum _{x \in \Omega }\, \frac{p(A \rightarrow \alpha )}{{P} (x, \Delta _{x})}\, \frac{\partial {P} (x, \Delta _{x})}{\partial \, p(A \rightarrow \alpha )}\,\right] _{\pi } \nonumber \\&\quad =\,h\,\left[ \,\prod _{x \in \Omega }\, {P} (x, \Delta _{x})^{h}\,\sum _{x \in \Omega } \frac{1}{{P} (x, \Delta _{x})} \sum _{t_{x} \in \Delta _{x}} N(A \rightarrow \alpha , t_{x}) {P} (x, t_{x})\,\right] _{\pi } \end{aligned}$$
(14)
Similarly, we will develop the expression
\(\,\sum _{i=1}^{n_{A}}\,{\mathcal {D}}_{A \rightarrow \alpha _i}^h (\Delta _x)\,\) as a preliminary step to evaluating the expressions
\(\,\sum _{i=1}^{n_{A}}\,{\mathcal {D}}_{A \rightarrow \alpha _i}^1 (\Delta _x^r)\,\) and
\(\,\sum _{i=1}^{n_{A}}\,{\mathcal {D}}_{A \rightarrow \alpha _i}^h (\Delta _x^c)\,\) of the denominator.
$$\begin{aligned}&\sum _{i=1}^{n_{A}}\; {\mathcal {D}}_{A \rightarrow \alpha _{i}}^{h} (\Delta _{x})\,=\nonumber \\&\;\;=\,\sum _{i=1}^{n_{A}}\;h\,\left[ \,\prod _{x \in \Omega }\, {P} (x, \Delta _{x})^{h}\;\sum _{x \in \Omega }\, \frac{1}{{P} (x, \Delta _{x})}\, \sum _{t_{x} \in \Delta _{x}}\,N(A \rightarrow \alpha _i, t_{x})\, {P} (x, t_{x})\,\right] _{\pi }\nonumber \\&\;\;=\,h\,\left[ \,\prod _{x \in \Omega }\, {P} (x, \Delta _{x})^{h}\;\sum _{x \in \Omega }\, \frac{1}{{P} (x, \Delta _{x})}\, \sum _{t_{x} \in \Delta _{x}}\;\sum _{i=1}^{n_{A}}\; N(A \rightarrow \alpha _i, t_{x})\, {P} (x, t_{x})\,\right] _{\pi }\nonumber \\&\;\;=\,h\left[ \,\prod _{x \in \Omega }\, {P} (x, \Delta _{x})^{h}\;\sum _{x \in \Omega }\, \frac{1}{{P} (x, \Delta _{x})}\, \sum _{t_{x} \in \Delta _{x}}\;N(A, t_{x})\, {P} (x, t_{x})\,\right] _{\pi } \end{aligned}$$
(15)
Given the expression
\(\,{\widetilde{F}}_h({\mathbb {G}}, \Omega )\,\) in Eq. (
7) and substituting expressions (
14) and (
15) in transformation (
13), we get the final expression after simplifying
\(\,\prod _{x \in \Omega }\,{P} (x, \Delta _{x}^{r})\) in the numerator and denominator.
$$\begin{aligned} {\bar{p}}(A \rightarrow \alpha )\,=\, \frac{{D}_{A \rightarrow \alpha } (\Delta _{x}^{r})\,-\,h\, {D}_{A \rightarrow \alpha } (\Delta _{x}^{c})\,+\, p(A \rightarrow \alpha )\; \frac{\textrm{C}}{\prod _{x \in \Omega }\,{P} (x, \Delta _{x}^{r})}}{{D}_{A} (\Delta _{x}^{r})\,-\,h\, {D}_{A} (\Delta _{x}^{c})\,+\, \frac{\textrm{C}}{\prod _{x \in \Omega }\,{P} (x, \Delta _{x}^{r})}} \end{aligned}$$
(16)
The auxiliary functions
\({D}_{A \rightarrow \alpha } (\Delta _{x})\) and
\({D}_{A} (\Delta _{x})\) will be given by
$$\begin{aligned} {D}_{A \rightarrow \alpha } (\Delta _{x})&=\;\sum _{x \in \Omega }\; \frac{1}{{P} (x, \Delta _{x})}\, \sum _{t_{x} \in \Delta _{x}}\;N(A \rightarrow \alpha , t_{x})\; {P} (x, t_{x})\,, \nonumber \\ {D}_{A} (\Delta _{x})&=\;\sum _{x \in \Omega }\; \frac{1}{{P} (x, \Delta _{x})}\, \sum _{t_{x} \in \Delta _{x}}\;N(A, t_{x})\,{P} (x, t_{x})\,. \end{aligned}$$
(17)
Gopalakrishnan et al. in [
16] suggest that to obtain a fast convergence and to guarantee the conditions of the growth transformations theorem for rational functions, the constant
\(\,\textrm{C}\,\) should be calculated as follows:
$$\begin{aligned} \textrm{C}\,=&{\text {max}}\,\left\{ \, \max _{p(A \rightarrow \alpha )}\,\left\{ \,-\, \frac{\partial \, {Q}_{\pi }({\mathbb {G}}, \Omega )}{\partial \, p(A \rightarrow \alpha )} \right\} _{\pi }, \, 0\,\right\} \,+\,\epsilon \end{aligned}$$
(18)
where
\(\epsilon\) is a small positive constant. Considering the expression
\(\,{Q}_{\pi }({\mathbb {G}}, \Omega )\,\) in Eq. (
8) and carrying out a development similar to the one we have done to obtain Eq. (
16), expression (
18) is as follows:
$$\begin{aligned} \frac{\partial \, {Q}_{\pi }({\mathbb {G}}, \Omega )}{\partial \, p(A \rightarrow \alpha )}\,=\, \frac{\prod _{x \in \Omega }\,{P} (x, \Delta _{x}^{r})}{p(A \rightarrow \alpha )}\;\left[ \,{D}_{A \rightarrow \alpha } (\Delta _{x}^{r})\,-\,h\, {D}_{A \rightarrow \alpha } (\Delta _{x}^{c})\,\right] \end{aligned}$$
Substituting this expression in Eq. (
18) allows us to calculate a
\(\textrm{C}\) maximum, (
\(\widetilde{\textrm{C}}\)), as:
$$\begin{aligned} \widetilde{\textrm{C}}\,=&{\text {max}}\,\left\{ \, \max _{p(A \rightarrow \alpha )}\,\left\{ \,-\, \frac{\left[ \,{D}_{A \rightarrow \alpha } (\Delta _{x}^{r})\,-\,h\, {D}_{A \rightarrow \alpha } (\Delta _{x}^{c})\,\right] }{p(A \rightarrow \alpha )}\; \right\} _{\pi },\,0\,\right\} \,+\,\epsilon \end{aligned}$$
Finally, expression (
16) becomes
$$\begin{aligned} {\bar{p}}(A \rightarrow \alpha )\,=\, \frac{{D}_{A \rightarrow \alpha } (\Delta _{x}^{r})\,-\,h\, {D}_{A \rightarrow \alpha } (\Delta _{x}^{c})\,+\, p(A \rightarrow \alpha )\;\, \widetilde{\textrm{C}}}{{D}_{A} (\Delta _{x}^{r})\,-\,h\, {D}_{A} (\Delta _{x}^{c})\,+\, \widetilde{\textrm{C}}}\,. \end{aligned}$$
This expression and auxiliary expressions (
17) coincide with expressions (
9), (
10), and (
11) as we had proposed.