r - constructing probability density distributions out of cumulative data of bins and counts - Cross Validated
we have set of data has been processed elsewhere , contains bins unequal size , counts of occurrence within each bin (around 8-10 bins, approximately log-like stretching). going analyze distribution in r. of statistical packages rather deal raw data, building histograms, deriving probability density distributions etc. not have raw, pre-counted data. how possible
- to construct in r probability density distribution function histogram (bar chart)?
- to perform kernel smoothing infer smoothed distribution function?
- to visualize cumulative function of total volume - again histograms and/or smoothed curve. in mathematica there histogramsmooth , allied functions.
the problem 1) not bare histogram height proportional counts (since bins unequal). particular application grain size analysis (or equivalent objects, example, distribution of oil fields worldwide on sizes). here example data:
bins: 60-100, 100-200, 200-300, 300-500, 500-1000, 1000-2000, 2000-3000, >3000 counts: 275, 320, 112, 65, 53, 44, 16, 15 bins ranges of size (volume), , counts numbers of objects of each range. special question refers plotting (and programming in r) such data, if last bin stated "above 3000" - how infer reasonable size , height visualization on histograms? then, need manually compute bins nonlinear scale on x, or package in r job including handling nonlinear scales?
Comments
Post a Comment