Data & Data TypesData & Data TypesStatistical models rely extensively on data to make predictions. In our case, the models are the statistical distributions and the data are the life data or times-to-failure data of our product. The accuracy of any prediction is directly proportional to the quality, accuracy and completeness of the supplied data. Good data, along with the appropriate model choice, usually results in good predictions. [Note: In using life data analysis (as well as general statistics), one must be very cautious in qualifying the data. The first and foremost assumption that must be satisfied is that our data, or our sample, is truly representative of the population of interest. Most statistical analysis assumes that the data are drawn at random from the population of interest. For example, if our job was to estimate the average life of a human being, we would expect our sample to have the same make-up as the general population, i.e. equal number of men and women, a representative percentage of smokers and non-smokers, etc. If in this case we used a sample of ten male smokers, the resulting analysis and prediction would most likely be biased and inaccurate. The assumption that our sample is truly representative of the population and that the test or use conditions are truly representative of the use conditions in the field must be satisfied in all analyses.] Bad, or insufficient data, will almost always result in bad predictions. (Note: This is adroitly summed up in the epigram "Garbage in, garbage out.")
In the analysis of life data, we want to use all available data which sometimes is incomplete or includes uncertainty as to when a failure occurred. To accomplish this, we separate life data into two categories: complete (all information is available) or censored (some of the information is missing). This chapter details these data classification methods.
This chapter is made up of the following sections:
Go to weibull.com
Go to ReliaSoft.com