predictr 0.1.5

Weibull Analysis Utilities

predictr

predictr: predict + reliability, in other words: A tool to predict the reliability.
The aim of this package is to provide state of the art tools for all kinds of Weibull analyses.
predictr already includes many methods (see tables below). A guideline when to use which method will be added soon.

Installation

Use the package manager pip to install predictr.

pip install predictr

Usage

Import predictr in python

from predictr import Analysis

How to use the Analysis class after importing

If both df and ds are None, an error will be raised. Configure all needed arguments that fits your use case.
Arguments to pass in Analysis class:
Analysis(df: list = None, ds: list = None, show: bool = False, plot_style='ggplot', bounds=None, bounds_type='2s', cl=0.9, bcm=None, bs_size=5000, est_type='median', unit='-'))

Step 1: Create an instance of the class Analysis
The class ibject is some prototype data, so we will call it "prototype_test". You can name it whatever you want.
prototype_test = Analysis(df = [200, 300, 400], show=True)

Step 2: Use either the mle() or mrr() method of the class Analysis
prototype_test.mle() -> MLE or prototype_test.mrr() -> MRR

Default Parameter values

df: list = None -> failures in seconds, days, no. of cycles etc., e.g. df = [100, 120, 80, 300]
ds: list = None -> suspensions (right-censored) in seconds, days, no. of cycles etc., e.g. ds = [300, 400, 400]
show: bool = False -> If True, the Weibull probability plot will be plotted.
plot_style = 'ggplot' -> Choose a style according to your needs. See https://matplotlib.org/3.1.0/gallery/style_sheets/style_sheets_reference.html for styles.
bounds=None -> Use following table to configure everything related to confidence bounds, e.g. if you want to use Monte-Carlo pivotal bounds for the Median Rank Regression: bounds = 'mcpb'.

confidence bounds	mle()	mrr()	uncensored data	censored data	type	argument value
Beta-Binomial Bounds	-	x	x	x	'2s', '1sl', '1su'	'bbb'
Monte-Carlo Pivotal Bounds	-	x	x	x	'2s', '1sl', '1su'	'mcpb'
Non-Parametric Bootstrap Bounds	x	x	x	-	'2s', '1sl', '1su'	'npbb'
Parametric Bootstrap Bounds	x	x	x	-	'2s', '1sl', '1su'	'pbb'
Fisher Bounds	x	-	x	x	'2s', '1sl', '1su'	'fisher'
Likelihood Ratio Bounds	x	-	x	x	'2s', '1sl', '1su'	'lrb'

bounds_type = '2s' -> '2s': two-sided confidence bounds, '1su': upper confidence bounds, '1sl': lower confidence bounds. E.g. bounds_type = '1sl'.
cl=0.9 -> configure the confidence level in the intervall (0, 1.0)
bcm=None -> Define the bias-correction method when the MLE is being used. Bootstrap bias-corrections are dependent on the number of bootstrap replication and the chosen statistic, e.g. if bcm = 'np_bs': bs_size = 5000 and est_type = 'median'.
bs_size = 5000 -> Resampling/Bootstrap sample size (number of replication). bs_size should be greater than or equal to 2000 for accurate results. The higher the number of replication, the longer it takes to compute the bias-correction.
est_type = 'median' -> When using bootstrap bias-corrections, this argument decides which statistic to compute from the bootstrap samples.
The following table provides possible configurations. Bias-corrections for mrr() are not supported, yet.

Bias-correction method	mle()	mrr()	argument value	config.	statistic
C4	x	-	'c4'	-	-
hrbu	x	-	'hrbu'	-	-
non-parametric Bootstrap correction	x	-	'np_bs'	bs_size	'mean', 'median', 'trimmed_mean'
Parametric Bootstrap correction	x	-	'p_bs'	bs_size	'mean', 'median', 'trimmed_mean'

unit = '-' -> Unit of the elements in df and ds, e.g. unit = 'seconds', unit = 'days', unit = 'ms' etc.

How to use the Maximum Likelihood Estimation (MLE)

Just add '.mle()' after Analysis()
object = Analysis().mle()

Uncensored sample

Example:

failures = [0.4508831,  0.68564703, 0.76826143, 0.88231395, 1.48287253, 1.62876357]
prototype_a = Analysis(df=failures, bounds='fisher',show=True).mle()

Censored sample

Example:

failures = [0.4508831,  0.68564703, 0.76826143, 0.88231395, 1.48287253, 1.62876357]
suspensions = [1.9, 2.0, 2.0]
prototype_a = Analysis(df=failures, ds=suspensions, bounds='lrb',show=True).mle()

How to use the Median Rank Regression (MRR)

Just add '.mrr()' after Analysis()
object = Analysis().mrr()

Uncensored sample

Example:

failures = [0.4508831,  0.68564703, 0.76826143, 0.88231395, 1.48287253, 1.62876357]
prototype_a = Analysis(df=failures, bounds='bbb',show=True).mrr()

Censored sample

Example:

failures = [0.4508831,  0.68564703, 0.76826143, 0.88231395, 1.48287253, 1.62876357]
suspensions = [1.9, 2.0, 2.0]
prototype_a = Analysis(df=failures, ds=suspensions, bounds='mcpb',show=True).mrr()

To Do

I will add a homepage with more detailed examples and guidelines for non-experts in the field of reliability engineering.