Struct PChart 

pub struct PChart { /* private fields */ }

Proportion nonconforming (P) chart.

Monitors the fraction of defective items in samples that may have different sizes. Control limits vary per subgroup when sample sizes differ.

Formulas

• CL = p-bar = total_defectives / total_inspected
• UCL_i = p-bar + 3 * sqrt(p-bar * (1 - p-bar) / n_i)
• LCL_i = max(0, p-bar - 3 * sqrt(p-bar * (1 - p-bar) / n_i))

Phase I and Phase II

PChart::new estimates p-bar from the samples it is given (Phase I). To judge later samples against an established process, build the chart with PChart::with_center: every limit then uses that p-bar with each sample’s own n. Re-estimating from Phase I and Phase II samples together would let a shift in Phase II pull the centre toward itself and widen its own limits.

Examples

use u_analytics::spc::PChart;

let mut chart = PChart::new();
chart.add_sample(3, 100).unwrap();  // 3 defectives out of 100
chart.add_sample(5, 100).unwrap();
chart.add_sample(2, 100).unwrap();
chart.add_sample(4, 100).unwrap();

let p_bar = chart.p_bar().expect("should have p_bar after adding samples");
assert!(p_bar > 0.0);

Reference

Montgomery, D.C. (2019). Introduction to Statistical Quality Control, 8th ed., Chapter 7, Section 7.3.

Implementations

impl PChart

pub fn new() -> Self

Create a new P chart that estimates p-bar from its samples (Phase I).

pub fn with_center(p_bar: f64) -> Result<Self, ControlChartError>

Create a P chart whose centre line is a known p-bar, typically from a Phase I study (Phase II).

Errors

ControlChartError::InvalidStandard unless p_bar is finite and strictly between 0 and 1 — at 0 or 1 every limit collapses onto the centre line.

pub fn add_sample( &mut self, defectives: u64, sample_size: u64, ) -> Result<(), ControlChartError>

Add a sample with the number of defective items and the total sample size.

Errors

ControlChartError::ZeroSampleSize if sample_size is 0, or ControlChartError::DefectivesExceedSampleSize if defectives > sample_size. The chart is left unchanged; such a sample used to be dropped silently, shifting every later point.

pub fn p_bar(&self) -> Option<f64>

The centre line in use — the known p-bar for a chart built with_center, otherwise the estimate — or None if no data.

pub fn points(&self) -> &[AttributeChartPoint]

Get all chart points.

pub fn is_in_control(&self) -> bool

Check if the process is in statistical control.

Trait Implementations

impl Default for PChart

fn default() -> Self

Returns the “default value” for a type.