Clarifying Data Analytics Concepts for Industrial Engineering

Clarifying Data Analytics Concepts for Industrial Engineering

1. Main concepts

   • Technical Algorithms
     • Artificial Intelligence
       • Branch of computer science
       • Performs functions associated with human intelligence
         • Reasoning
         • Learning
         • self--improvement
     • Machine Learning
       • Automatics learning, process by which a unit improves its performance by acquiring new knowledge or skills
         • Both fields connected because of data analysis algorithms
   • Models
     • Bid Data
       • Extensive datasets that require efficient storage, manipulation and analysis
       • Characteristics
         • Volume
         • Variety
         • Velocity
         • Variability
     • Data Analytics
       • Initial standard statistical techniques
         • Construct initial example set
       • Machine learning algorithms
         • Validate example set
     • Data Mining
       • Extraction of implicit, previously uknown and useful information from data

2. Extraction of value from data

   • Manufacturing industry generates 1/3 of all world data
   • Data needs to be analysed
   • Data analytics for manufacturing companies to stay competitive
   • Data mining and machine learning as ways to collect data and achieve competitive advantages

3. Machine learning techniques in smart manufacturing

   • Supervised Learning
     • Linear Methods for Regression
       • Usage
         • Provide an adequate and interpretable description of how the inputs affect the output
       • Applications
         • Process monitoring
         • Process control
         • Quality prediction
     • Linear Method for Classification
       • Uses the variable's characteristicss to identify which class the variable belongs to
     • Decision Tree Based Method
       • "Divide-and-conquer" approach to the problem of learning from a set of independent instances
     • Neural Networks
       • Extract linear combinations of the inputs as new features and then model the target output as a nonlinear function of these features
       • Applications
         • Manufacturing process control
         • Machine fault diagnosis
   • Unsupervised Learning
     • Principal Component Analysis
       • Converts a set of observations of possibly correlated variables into a set of values of linearly uncorrelated variables
       • Applications
         • Process monitoring
         • Product quality
         • Machine fault detection
     • Association Rules
       • Find joint values of the input variables that appear most freequently in the dataset
       • Applications
         • Predict final product quality
         • Analyse causes of product defects
     • Cluster Analysis
       • Recognizes whether a dataset consists of subgroups, each group representing objects with different properties
       • Applications
         • Machine fault detection
         • Process monitoring
     • Self-Organizing Maps
       • Method consisting of nodes that are linked with a weight vector and a position in the map space

4. Conclusion

   • Extracting value from data is paramount
   • Summary of the tecniques and cconcepts that are currently used in Industrial engineering practice