MLE_stage2_gradient: Stage 2: solve likelihood directly
In adzemski/netprobitFE: Network probit with fixed effects

Stage 2: solve likelihood directly

1
2
3

MLE_stage2_gradient(wide, bound_lower, bound_upper, rho_start = 0,
  inital_search_radius = 0.4,
  convergence_tol = sqrt(.Machine$double.eps))

`wide`	network data in upper-triangular wide format
`bound_lower`	lower bound on rho
`bound_upper`	upper bound on rho
`rho_start`	starting value (default = 0)
`inital_search_radius`	intial search radius around starting value (default = 0.4)
`convergence_tol`	convergence tolerance