Control Chart for process learnining: CEP flexibility with Artificial Intelligence (AI)
Artificial Inteligence, CEP, Control Charts, Classification Models, Regression Models, Resampling.
In this work, we use Artificial Intelligence and Statistical Process Control techniques in two real data sets from the civil construction area, in the presence of covariates and continuous and categorical observations. The data represent the daily ceramic laying process under observation of interest (customer satisfaction levels for the first data set and process cost for the second data set) in four explanatory variables to this process. For process monitoring, based on the Multiple Regression Control Chart proposed by Haworth (1996), the traditional p-Chart for attribute control, and some control tools that use artificial intelligence, two new types of graphs control are developed. Called the Classification Control Chart and Prediction Control Chart for Continuous Responses, its development was based on concepts of statistical modeling (parametric and nonparametric), as well as resampling processes using machine learning and artificial intelligence. Firstly, some classification and regression models, competitive in their predictive capacity, are presented in such a way that the selected modeling results in estimates as close as possible to the real values. For this, using machine learning techniques, performance metrics are compared. Next, resampling procedures, such as Cross-Validation k- textit fold, are proposed in order to ensure maximum information extraction from the data that will be used to construct the control limits of the graphs. Simulation studies were performed with the objective of comparing, based on the Average Run Length (ARL), the process monitoring performance of the proposed graphs in different simulated scenarios.