Introduction
MIPAR is able to fit a probability distribution curve to a measurement histogram. The supported fit types are: Normal, Log-normal and Bimodal.
Normal Fit
MIPAR uses a Gaussian distribution function to generate a probability curve from the feature measurements.
Equation describing the curve:
MIPAR reports the mean, standard deviation, and mode using the mu-1 and sig-1 parameters, to display parameters select ‘Show Fit Parameters’
Log-normal Fit
MIPAR uses a Galton distribution function to generate a probability curve from the feature measurements:
Equation describing the curve:
MIPAR reports the mean, standard deviation, and mode using the mu-1 and sig-1 parameters, to display parameters select ‘Show Fit Parameters’.
Note that mu-1 and sig-1 are the mean and standard deviation of logs and therefore describe the curve and not the mean and standard deviation in the original measurement units. To calculate the log normal fit peak (mode) use this equation:
Bimodal Fit
MIPAR uses a Gaussian mixture model with two peaks to generate a fit from the feature measurements. MIPAR reports the mean, standard deviation, mode, and mixing proportion of each component using the mu-1,2; sig-1,2; mix-1,2 parameters, to display parameters select ‘Show Fit Parameters’.
Measurement at Percentile
MIPAR reports the inverse cumulative distribution function values of the Normal and Log-normal fits at probabilities 10%, 25%, 75%, 90%, noted as D10, D25, D75 and D90.
Need more help with this?
Chat with an expert now ››
