WORLD PULP&PAPER 97 Without regular tuning, today’s predictive models quickly become irrelevant in highly variable applications such as papermaking. seasonal changes, operating crew and upstream operations are among the many sources of variability in paper manufacturing. While some of these sources of variability may be directly captured in the predictive model, most are not. Without regular tuning, today’s predictive models quickly become irrelevant in highly variable applications such as papermaking. While manually tuning of predictive models is an option, the process is time-consuming, costly, inefficient, and requires a subject matter expert. Current predictive models have a number of disadvantages. For example: • They require full-time dedication of mill resources for development, monitoring, and interpretation. • They are expensive to develop. • They have a lengthy implementation process. • They are not developed specifically for the papermaking market. • They are not supported by an onsite representative. • They require manual tuning. • They focus primarily on asset protection and preventative maintenance. While each product offering presents a unique method for applying advanced analytics in the papermaking process, all lack the ability to provide predictions of key quality measures that remain robust and accurate in a continuous, time- stamped manufacturing environment. THE SOLENIS SOLUTION OPTIX Applied Intelligence is a unique adaptive analytics platform built with the latest artificial intelligence (AI) and machine learning capabilities available today. OPTIX was developed Figure 2. (a) Comprehensible relationships between strength, basis weight, and smoothness; (b) Causal network showing multidimensional relationships between strength, basis weight, smoothness, and numerous other predictors.