#LATIN HYPERCUBE SAMPLING SCRIPT PYTHON INSTALL#

#LATIN HYPERCUBE SAMPLING SCRIPT PYTHON INSTALL#

YouвЂ™ll probably need to install SALib if itвЂ™s not a package youвЂ™ve used already. To perform the sensitivity analysis, I am going to use SALibВ (featured in this blog multiple times already), to perform a Delta Moment-Independent Analysis (also produces a first order Sobol sensitivity index ). In other words, if some factors appear to be driving the frequent small shortages experienced, are those factors the same for the rare large shortages? For this particular example, we aim to see what are the most significant drivers of shortage at the different levels itвЂ™s experienced by this user. The analysis will be magnitude-varying, i.e., it will be performed at different magnitudes of our output of interest. In this post IвЂ™m going to walk the reader through performing a sensitivity analysis using the output of an experiment using multiple Latin Hypercube Samples. These might seem obvious to the readers of this blog, but often times we perform our analyses for only some of these percentiles, вЂњthe worst eventвЂќ, вЂњthe averageвЂќ, etc., which is certainly very informative, but can potentially miss part of the bigger picture. The various percentiles tell a different part of the story of how a system operates, the 5th percentile tells as that its level is exceeded 95% of the time, the 99th tells as that its level is only reached once in every 100 years in our records. вЂњ how many years exceed an annual shortage of 1000 af?вЂњ).

вЂњhow bad is the worst case?вЂќ, вЂњhow rare is the worst case?вЂќ), or we might like to know how often we exceed a certain threshold (e.g. range of experiment outputsДЄs already discussed, there are multiple benefits for visualizing the output in such manner: we are often concerned with the levels and frequencies of extremes when making decisions about systems (e.g.

Latin hypercube sampling script python