Transcription
Data Analytics with MATLABDr. Jan EggersMathWorksJune 9, 2015 2015 The MathWorks, Inc.1
From Data to Decisions & ions &DesignKnowledgeAction20NNmeasured1200.910active power per-unitWeightDisplacement 020Acceleration200400 2000Displacement4000Weight50 100150200Horsepow er20406080100120time secs1401601802002
Key Takeaways Access and preprocess large amounts of heterogeneous data Develop advanced analytics with machine learning Integrate analytics with your enterprise systems3
4
Big Data Capabilities in MATLABMemory and Data Access 64-bit processors Memory Mapped Variables Disk Variables Databases DatastoresProgramming Constructs Streaming Block Processing Parallel-for loops GPU Arrays SPMD and Distributed Arrays MapReducePlatforms Desktop (Multicore, GPU) Clusters Cloud Computing (MDCS on EC2) Hadoop5
DataStoredatastoreImport text files & collections of text filesthat don’t fit into memoryds datastore('file1.mat');ds datastore('*.csv');ds datastore('/shared/data repository/');ds s datastore({'/shared01/','/shared02/'});while hasdata(ds)T read(ds);end6
MapReduceData 14DL914US245maxDelay mapreduce(ds, @maxDistMapper, @maxDistReducer);7
A Big Data educeMapMapReduceReduceMapMapReduceReduceDataData8
9
Machine LearningMachine learning uses data and produces a program to perform a taskTask: Human Activity DetectionMachine Learning ApproachStandard ApproachMachineLearningComputerProgramHand Written ProgramIf X acc 0.5then “SITTING”If Y acc 4 and Z acc 5then “STANDING” Formula or Equation𝑌𝑎𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝛽1 𝑋𝑎𝑐𝑐 𝛽2 𝑌𝑎𝑐𝑐 𝛽3 𝑍𝑎𝑐𝑐 𝑚𝑜𝑑𝑒𝑙: Inputs ���𝑚𝑜𝑑𝑒𝑙 (𝑠𝑒𝑛𝑠𝑜𝑟 𝑑𝑎𝑡𝑎, 𝑎𝑐𝑡𝑖𝑣𝑖𝑡𝑦)10
Machine Learning TechniquesMachineLearningType of LearningCategories of AlgorithmsUnsupervisedLearningClusteringGroup and interpretdata based onlyon input dataSupervisedLearningClassificationDevelop predictivemodel based on bothinput and output dataRegression11
Apply Machine Learning techniques easilyMachineLearningData: 3-axial Accelerometer data 3-axial Gyroscope data12
Data Analytics WorkflowAccessExploreStart locally Share/Deploy Start “simple”Basic statisticsExplore dataWork on your desktop13
Data Analytics WorkflowPrototypeAccessExplore prototype algorithms and then .Share/Deploy Interactive developmentAdvanced algorithmsWork on your desktop14
Data Analytics WorkflowPrototypeAccessExploreShare/DeployScale scale up as needed Parallel Computing Tools Scale to a cluster15
16
A Primer on Deploying MATLAB Programs Excel add-insDesktopWeb &Enterprise Royalty-free Encryption to protect intellectual propertyMATLABProductionServer(s)WebServer(s)17
Benefits of Deploying MATLAB Code Domain experts maintain ownership of ideas, algorithms, and applications Flexibility to integrate with different programming languages Implement a common algorithm on different platforms Avoid time consuming and error prone re-coding Easily adopt algorithm improvements throughout lifecycle18
Predictive Data Analytics – Load Demand Forecasting19
Big Data and Predictive Analytics at Shell20
STIWA Increases Total Production Output of Automation MachineryChallengeApply sophisticated mathematical methods to optimizeautomation machinery and increase total production outputSolutionUse AMS ZPoint-CI to collect large production data sets innear real time and use MATLAB to analyze the data andidentify optimal trajectoriesResults Total cycle time reduced by 30% Large data sets analyzed in seconds Deployment to multiple machines streamlinedSTIWA’s shopfloor managementsystem, based on MATLAB, AMSZPoint-CI, and AMS Analysis-CI.“Our shopfloor management system AMSZPoint-CI collects a huge amount ofmachine, process, and product data 24 hoursa day. By analyzing this data immediately inMATLAB and AMS Analysis-CI we haveachieved a tenfold increase in precision, a30% reduction in total cycle time, and asignificant increase in production output.”Alexander MeisingerLink to user storySTIWA21
Key Takeaways Access and preprocess large amounts of heterogeneous data– Capabilities to deal with big data are available and evolve– Tools to organize data and automate the process Develop advanced analytics with machine learning– Advanced statistical and machine learning methods to gain insights– Apps to rapidly iterate through and assess different models Integrate analytics with your enterprise systems– Parallel Computing and Map Reduce to scale up as needed– Application deployment on every scale to make models available to others22
Data Analytics with MATLAB Dr. Jan Eggers MathWorks June 9, 2015. 2 MPG Acceleration Displacement Weight Horsepower G n t t r 20 40 10 20 200 400 2000 4000 50 100150200 50 100 150 200 2000 4000 200 400 10 20 20 40 From Data to Decisions & Design . Integrate analytics with your enterprise systems. 4 Agenda Data Techniques Goal Explore .
