Sen Slope estimator

Non-parametric Slope Using the Method of Theil (1950) and Sen (1968)

%% INPUTS:

% TSD = (n x 2) double

% TSD(:,1) = Time, Eg. years, TSD(:,1) = [1991; 1992; 1993; ... 2015]

% TSD(:,2) = Data, Eg. Data may be the average temprature of the years

% alpha = significance level of the test, Eg. alpha = 0.05

% Plot = 0 or 1; (Plot = 1 : Plots the time series data and Sen's

% slope estimators)

%% OUTPUTS:

% H = hypothesis test results (H = 1 : Trend is significant)

% beta1 = Sen's slope

% LL = Lower confidence interval for Sen's slope

% UL = Upper confidence interval for Sen's slope

% LL < beta1 < UL

% beta0 = Intercept

% H05 = hypothesis test results for beta1/slope at the 5%

% (alpha = 0.05) significance level

% H01 = hypothesis test results for beta1/slope at the 1%

% (alpha = 0.01) significance level

%% RUNNING SAMPLE:

% >> load('tsd_sample.mat');

% >> Sen_Slope(tsd_sample, .05, 1);