The decay of Six Sigma and the rise of Quality 4.0 in manufacturing innovation

Figures & data

Table 1. Acronym definitions.

Acronym	Definition
AA	Advanced analytics
ANN	Artificial neural networks
AI	Artificial intelligence
BD	Big data
CSC	Cloud storage and computing
DL	Deep learning
DMAIC	Define, measure, analyze, improve, and control
DPMO	Defects per million of opportunities
IBD	Industrial big data
IIoT	Industrial internet of things
IM	Intelligent manufacturing
I4.0	Industry 4.0
ML	Machine learning
MLA	Machine learning algorithms
PCA	Principal components analysis
PMQ	Process monitoring for quality
Q4.0	Quality 4.0
SM	Smart manufacturing
SPC	Statistical process control
SQC	Statistical quality control
TQM	Total quality management

Table

Cause-and-effect diagram	Response surface methodology (RSM)
Check sheet	Analysis Of Variance (ANOVA)
Control chart	General linear model
Histogram	ANOVA Gauge R
Pareto chart	Regression analysis
Scatter diagram	Correlation
Stratification	χ^2 test

Figure 1. Google Trends, interest of Six Sigma from 2004 to 2022. Interest over time. The numbers represent the search interest relative to the highest point on the chart for the given region and time.

Figure 2. Areas of knowledge of Q4.0.

Table 2. Confusion matrix.

	Predicted good	Predicted bad
Good item	True negative (TN)	False positive (FP)
Defective item	False negative (FN)	True positive (TP)

Figure 3. PMQ scheme.

Figure 4. PMQ problem solving strategy (Escobar, Macias et al. 2022).

Table 3. Quality 4.0 certifications curricula.

Certifications
Yellow	Green	Black	Master black
Smart manufacturing	Binary classification of quality	Data preprocessing techniques	Deep learning
-Definitions	-Machine learning theory	-Visualization	-Feed forward neural network
-Paradigms	-Machine learning algorithms	-Missing values	-Convolutional neural network
-Ciber physical systems	Applications	-Imputation methods	Meta-learning
Manufacturing big data	Problem solving strategy	-Feature scaling	-Multiple classifier systems
−10 V’s	-Identity	-Outlier analysis	Managerial implications
-Data sets for binary classification of quality	-Acsensorize	-One hot encoding	-Team building
Traditional Quality Control Methods	-Discover	Feature Selection	-Project selection theory
-Definitions	-Learn	-Filter methods	-Outsourcing evaluation
-Advantages	-Predict	-Wrappers
-Limitations	-Redesign	-Embedded
Quality 4.0	-Relearn	Model Selection
-Definitions		-Parsimonious modeling
-Research areas
-Advantages
-Limitations
Technologies
-Artificial Intelligence
-Cloud Storage and Computing
-Industrial Internet of Things

Figure 5. Hierarchy by Q4.0 belts.

Table 4. Six Sigma versus Quality 4.0.

	Six Sigma	Quality 4.0
Contrastive dimensions	Statistical-based	Machine learning-based
Data	Experimental	Observational
Feature space	Small to moderate	Big (hyper-dimensional)
Learning	Hypothesis-based	Discovery-based
Model	Explanatory	Predictive
Assumptions	Theoretical Prob Dist, homoscedasticity	No
Validation	Statistical significance (p value)	Generalization
Execution	Real tie and asynchronous	Real-time
Data Snooping	Not allowed	Yes (iterative modeling)
Solution	Long term-experimentally derived	Short term-observational data derived
Relearning	No	Yes (adaptive modeling)
Data partition	No (all data is used for model development)	Yes (training, validation, test)

Figure 6. Kaggle winners.

Figure 7. Kaggle winners.

Figure 8. Data driven and ML methods supporting engineering knowledge discovery.

Figure 9. Year-over-year percentage of change in publications by quality philosophy.

Figure 10. 3D quality pattern. Green triangles denote good quality, and red circles denote a defective item.

Table 5. Virtual feature values and the associated quality status of each sample.

Index	Feature 1	Feature 2	Feature 3	Quality status
1	89	91	95	Defective
2	33	20	98	Good
3	95	93	88	Defective
4	40	45	56	Good
5	80	86	92	Defective
6	20	55	12	Good
7	52	60	80	Good
8	90	95	20	Good
9	99	89	60	Good
10	85	82	75	Defective

Figure 11. Decision tree for choosing a control chart (Minitab).

Figure 12. Control charts: I-MR univariate control chart – Feature 1.

Figure 13. Control charts: I-MR univariate control chart – Feature 2.

Figure 14. Control charts: I-MR univariate control chart – Feature 3.

Figure 15. Control charts: Multivariate Hotelling’s T² – Features 1–3.

Figure 16. Control charts: Multivariate generalized variance control charts – Feature 1–3.

Figure 17. Linear separation scheme based on the support vector machine.

Table 6. Confusion matrix for this case study using SVM.

	Predicted good	Predicted bad
Good item	4	0
Defective item	0	6

Escobar, C., D. Macias, M. Mcgovern, M. Hernandez de Menendez, and R. Morales-Menendez. 2022. Quality 4.0 – An evolution of Six Sigma DMAIC. International Journal of Lean Six Sigma 5:1200–38.

 Google Scholar