mladenjovanovic/STM: Quantile regression and weighting

Added different weighting options for the estimate_ family of functions. These include

• "none" (for equal weight, or no weighting of the observations)
• "reps" (for 1/reps weighting)
• "load" (for using weight or %1RM)
• "eRIR" (for 1/(eRIR+1) weighting)
• "reps x load"
• "reps x eRIR"
• "load x eRIR"
• "reps x load x eRIR"

Added strength_training_log dataset. Single individual performing two strength training sessions per week, over the course of 12 weeks (4 phases, each 3 weeks long). Individual eRIR (estimated reps-in-reserve) subjective rating is included in the dataset. This dataset is used to demonstrate techniques for embedded testing of the 1RM and individual profiles

Added estimate_k_quantile(), estimate_kmod_quantile(), and estimate_klin_quantile() functions to implement non-linear quantile estimation of the parameters