In this article, a single parametric class of
modifications for Kovarik's method is proposed. It is proved that
all methods in this class are quadratically convergent. Numerical
comparison among methods of Kovarik, Petcu-Popa [5], and a special
method in this class, chosen based on a specific value for the
parameter, shows that Kovarik and Petcu-Popa's methods give almost
similar convergence results. However, the special method
converges faster and its iteration number is considerably lower
than that of others. For Numerical experiments, there are used ten
$n\times n$ test matrices with $n=5,10,20,50$, whose condition
numbers vary in the interval [$2\,,\,8.1e146$].