[ -- Internship Proposals -- ]

OpenCPU | Chart.js | Michaelis Menten

function(x, y) { nearest <- which.min(abs(y - 0.5 * max(y))) # index of Vmax/2 nlmodel <- nls(y ~ A * x / (B + x), start = list(A = max(y), B = x[nearest])) # Get summary and coefficients fit_res <- summary(nlmodel)$coefficients A_est <- fit_res[1, 1] # Estimate of A B_est <- fit_res[2, 1] # Estimate of B # Standard errors A_se <- fit_res[1, 2] # Standard error of A B_se <- fit_res[2, 2] # Standard error of B # Calculate degrees of freedom n <- length(y) # number of observations p <- length(coef(nlmodel)) # number of parameters (A and B) df <- n - p # degrees of freedom # Calculate critical t-value for 68% CI t_val <- qt(0.84, df) # For 68% CI, use 0.84 (1 - 0.32/2) # Calculate margin of errors A_margin_of_error <- t_val * A_se B_margin_of_error <- t_val * B_se # Extra values for a smooth display (not needed when # points is large) S_xtra <- c(seq(0, max(x) * 1.2, length.out = 49)) # create a dummy S (x) vector (49 values here) vfit_xtra <- A_est * S_xtra / (B_est + S_xtra) # calculate new v (y) values according to those 49 new values # Calculate RMSE error_squared <- (predict(nlmodel) - y) ^ 2 RMSE <- sqrt(mean(error_squared)) # Return the results return(c(A_est, B_est, A_margin_of_error, B_margin_of_error, RMSE, S_xtra, vfit_xtra)) }

Code (R)

Info

🙅 🙅

R-squared [1] | R-squared [2] | RMSE