I am running the following code:
ggplot(data= data_nickel_t, aes( x=index(data_nickel_t), y= log(ni_demand) )) +
scale_x_yearqtr(format = "%Y-%q", n = 14) +
geom_point() + stat_summary(fun.data=mean_cl_normal) +
geom_smooth(method='lm', aes(colour = "linear fit"), se= FALSE) +
geom_smooth(method='lm', formula = y ~ x + poly(x, 2), size = 1, aes(colour = "quadratic"), se= FALSE) +
geom_smooth(method='lm', formula = y ~ x + poly(x, 3), size = 1, aes(colour = "polynomial"), se= FALSE ) +
ggtitle("Global Refined Nickel Demand") +
xlab("Time") +
ylab("Thousand Metric Tons")
The code above produce a graph with three fitted lines but I get the following warning messages:
1: In predict.lm(model, newdata = data.frame(x = xseq), se.fit = se,
prediction from a rank-deficient fit may be misleading;
2: In predict.lm(model, newdata = data.frame(x = xseq), se.fit = se, :
prediction from a rank-deficient fit may be misleading;
3: Removed 94 rows containing missing values (geom_pointrange).
My first impression was collinearity between time trends variable in poly() function. I might estimate numerical model to check this further. As for the missing value issue, e.g this link explain the reasons for missing k rows. When I tried solutions suggested in that link, it does not work in my case, I still get the same error. I have 94 observations. I also don't have zeros in my data so no reason for log transformation to drop my values. I am still kind of new using r with time series any idea how I may fix the missing value warning?
