r - Loop linear regrssion model -
i have data amount dependent variable , len,age, quantity , pos explanotry variables. trying make regression of amount on age, quantity , pos using stepwise.
id sym month amount len age quantity pos 11 10 1 500 5 17 0 12 22 10 1 300 6 11 0 11 33 10 1 200 2 10 0 10 44 10 1 100 2 11 0 11 55 10 1 150 4 15 0 12 66 10 1 250 4 16 0 14 11 20 1 500 5 17 0 12 22 20 1 300 6 11 0 11 33 20 1 200 2 10 0 10 44 20 1 100 2 11 0 11 55 20 1 150 4 15 0 12 66 20 1 250 4 16 0 14 77 20 1 700 4 17 0 11 88 20 1 100 2 16 0 12 11 30 1 500 5 17 0 12 22 30 1 300 6 11 0 11 33 30 1 200 2 10 0 10 44 30 1 100 2 11 0 11 55 30 1 150 4 15 0 12 66 30 1 250 4 16 0 14 11 40 1 500 5 17 2000 12 22 40 1 300 6 11 1000 11 33 40 1 200 2 10 1000 10 44 40 1 100 2 11 1000 11 55 40 1 150 4 15 1000 12 66 40 1 250 4 16 1000 14 and output of results want after running regression should dataframe that's (that should me detect outliers):
id month sym amount len age quantity pos r^2 cookdistanse residuals upperlimit lowerlimit 11 1 10 500 5 17 null 12 0.7 1.5 -350 -500 1000 22 1 10 300 6 11 null 11 0.8 1.7 -400 -500 1000 that's code trying run on sym = 10, sym= 20, sym = 30, sym = 40. have 400 sym values run regression analysis on them.
fit[i] <- step(lm (sym[i]$sum ~ len + age + quantity, na.action=na.omit), direction="backward") r_sq <- summary(fit[i])$r.squared res[i] <- resid(fit[i]) d[i] <- cooks.distance(fit[i]) q[i] <- quantile(resid(fit[i), c(.25, .50, .75, .99)) l[i]<- q[1][i] - 2.2* (q[3][i]-q[1][i]) u[i] <- q[3][i] + 2.2*(q[3][i]-q[1][i]) "i" means results regression of sym = (10,20..). way on loop every sym value? highly appreciate.
Comments
Post a Comment