Documentation
< All Topics
Print

3.3 Model Training

Traditionally, we split data into training and validation sets in 80:20 ratio. Before we start the model training, make sure you have ingested data into Momentum and created training and validation sets. If you have not done so, here are links that will helps you to prepare the data. 

How to ingest data Momentum 

How to create training and test sets using Transformer 

How to create a data processing pipeline 

To explain the process of training a model, we will use deep learning or artificial neural network based multi-layer perceptron classifier that predicts if a machine will fail given its operating condition. We have ingested the AI4I 2020 Predictive Maintenance Dataset (https://archive.ics.uci.edu/ml/datasets/AI4I+2020+Predictive+Maintenance+Dataset), and created 80% training and 20% test sets.  

To train a model, e.g., multi-layer perceptron classifier, here are the steps: 

  1. Expand “ML Model” from under the “Machine Learning” section of the left side menu panel, and click “ML Home” to launch the ML home page. 
  1. Click “Create New Model”.  
  1. From the “Supervised Learning: Classification” drop down, select “Deep Learning/ Artificial Neural Network/Multilayer Perceptron Classifier.  
  1. Fill out the form as described below, and shown in Figure 3.1a and 3.1b for example: 
  1. Model Name: give a meaningful name to identify this model, for example Machine_Failure_Prediction_Model 
  1. Give a version number, just in case you need to build multiple versions of this model. 
  1. In the Feature Field text area, supply a comma separated list of all features you want the model to learn from. In this example, we are using the following features: 

Air_temperature,Process_temperature,Rotational_speed,Torque,Tool_wear_in_min 

Listing 3.1: Feature list 

  1. Categorical/Non-numerical fields: comma separated list of all features that are not numeric. We will not have such field, so we will leave it empty. 
  1. OneHot Encodable Field: Categorical and non-numeric fields should be encoded. Leave it empty in this case. 
  1. Target Field that needs to be predicted. Machine_failure is our target field. 
  1. Number of Classes: The machine failure data has only two classes – 0 means no failure and 1 means failure. We will fill 2 in this field. 
  1. Scale Features: It is generally a good idea to scale the features, we will select “yes”. 
  1. Feature excluded from scaling: We will leave this field empty. 
  1. Number of hidden layers: This is to configure the neural network. We will start with 3 hidden layers. 
  1. Number of nodes in each hidden layer: We will use varying number of nodes in each layer. For example, 19,11,9 to indicate we want to use 19 nodes in the first hidden layer, 11 in the second and 9 in the last hidden layer. If you want to use the same number of neurons in each hidden layer, use a single number. For example, if we enter 16 in this field, all hidden layers will have 16 neurons. 
  1. Max Number of Iteration: We are starting with 1000. If the algorithm converges before it reaches the max 1000 iterations, the training will automatically stop to avoid unnecessary computation and time.  
  1. Training: Test ratio to split the training data internally into this ratio. We are using 0.8:0.2 for 80% and 20% split. 
  1. Max Core: This is to parallelize the training by using multiple CPU cores of the cluster. We are going to use 16 cores as our dataset is not large. For large dataset, using more or all available CPU cores of the cluster will speed up the training process. 
  1. Memory per Core: For most training 4GB per core should be sufficient but may be increased for a large and complex model. 
  1. The number of partitions is not used at this time and is reserved for the future. 

Figure 3.1a: Showing a part of the neural network configuration form. 

Figure 3.1b: Showing the remaining part of the neural networks form 

  1. After submitting the form, a rectangular config widget is created on the main body of the page. 
  1. Expand Transformer from the left menu panel and click on the training set transformer to bring it to the main page. 
  1. Click on the “Out” on the training set rectangle and click on the “In” of the model config rectangle. This will join the training dataset to the model config (shown in Figure 12 below) 
  1. Saving the configuration will take the screen to the ML home page. 
  1. Click “Run” located at the top menu to start the model training. 

Figure 3.2: Screen showing model configuration 

Table of Contents

Lester Firstenberger

Lester is recognized nationally as a regulatory attorney and expert in consumer finance, securitization, mortgage, and banking law.

Lester is recognized nationally as a regulatory attorney and expert in consumer finance, securitization, mortgage, and banking law. In a variety of capacities, over the past 30 years as an attorney, Mr. Firstenberger has represented the interests of numerous financial institutions in transactions valued in excess of one trillion dollars. He was appointed to and served a three-year term as a member of the Consumer Advisory Council of the Board of Governors of the Federal Reserve System. He has extensive governmental relations experience in the US and Canada at both the federal and state and provincial levels.

Shamshad (Sam) Ansari

Shamshad (Sam) Ansari is the founder, president and CEO of Accure. He drives technology innovations and works with a great team of engineers, data scientists, and business drivers at Accure.

Shamshad (Sam) Ansari is the founder, president, and CEO of Accure. He drives technology innovations and works with a great team of engineers, data scientists, and business drivers at Accure. He takes great pride in working with customers and putting together teams for solving their business problems. Sam is the product architect of Momentum, an AI and automation platform for data engineers, scientists, and business analysts.

Sam brings more than 20 years of technology development and management expertise. He developed, deployed and managed several large scale AI projects. He is a domain expert in healthcare systems, protocols, standards and compliances. Sam is a serial entrepreneur and worked with 4 startups. Prior to starting Accure, he worked with Apixio as the principal architect and director of engineering. He had another successful startup Orbit Solutions where he developed healthcare systems that went through an acquisition. He worked with IBM and the US Government at various capacities.

Sam is a distinguished data scientist, inventor and author. He has several technology publications in his name. He has co-authored 4 US Patents in healthcare AI. He is a well respected authority in computer vision and AI and has authored a book, “Building Computer Vision Applications Using Artificial Neural Networks” that is also translated into other languages including Chinese. Sam contributes to academia as well. He mentors graduate students and sponsors Capstone projects. He is also a member of the Advisory Board, Data Analytics Engineering Department at George Mason University.

Sam has a Master’s degree from Indian Institute of Information Technology & Management, Kerala (IIITM-K) and Bachelor’s degree in engineering from Bihar Institute of Technology Sindri (BIT Sindri).

Moghisuddin Raza

Mogishuddin Raza is a technology leader. As the COO of Accure he is having global product delivery responsibility along with overall strategic and operational responsibility.

Mogishuddin Raza is a technology leader. As the COO of Accure he is having global product delivery responsibility along with overall strategic and operational responsibility.

Having extensive background in technology product development and integration, in particular to Enterprise storage, virtualization, cloud computing, high availability & business continuity technology/solutions, and Big Data & related technologies. Has been passionate and evangelizing the usage of Big data technologies using Momentum to implement advanced analytics (descriptive and predictive) to directly impact the business via an intuitive set of use cases.

Having approximately two decades of experience in high-tech industries which includes big MNCs corporate like EMC Corp and Hewlett-Packard to mid-size organization such as Netkraft, Trados Inc driving transformation in strategizing, planning and architecting product engineering, execution and delivery of high quality products releases within budget & time.

Skilled in all aspects of big MNCs as well as company startups and growth including: strategizing, business planning, market research, finance, product development and profit margins & revenue management. Excellent leadership and people motivation skills. Expert in managing cross-functional, cross cultural global team and building strategic partnership in the global virtual matrix team environment.

Overall, a senior software business professional, skilled in the management of people, resources and partnerships which enables building an eco system for a winning organization.



Rajesh Kumar Nedungadi

Scion of A Former Royal House of Kerala, India President Garuttman Group, USA. Rajesh is an entrepreneur & visionary specializing in International Business Strategy and Market Development with focus on Middle East & North America. With over 20 years’ experience in international trade, Rajesh is an expert on Business Strategy Development, Market Opportunity Development and International Sales & Marketing of Products and Services including the IT Industry. Rajesh is working as Managing Partner / Board Member of many companies including, Globistic Company USA, Castlewick Companies, USA.