Build and deploy your machine learning models effortlessly (2)
- 1. Developer Series Join our webinar series to expand your developer skills! Security Tuesdays Data Science and AI Wednesdays Cloud Native and Red Hat OpenShift Thursdays meetup.com/IBM-Cloud-MEA
- 2. Build and deploy your Machine Learning models effortlessly — Sidra Ahmed Developer Advocate, UAE Anam Mahmood Developer Advocate, UAE Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 3. Let’s get started • Sign up/Log in to your IBM Cloud Account https://ibm.biz/BdfEMZ • Follow along for the hands-on: https://github.com/Anam- Mahmood/watson-studio-learning-path- assets Get started at: https://ibm.biz/BdfEMZ Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation
- 4. 4 chat with everyone! Q&A here! Follow us to get notified about upcoming events View more info about this event Workshop Resources Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation
- 5. Survey https://www.surveygizmo.com/s3/6083679/9e93461cbf3e Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 6. Agenda 6 Jupyter Notebooks 16 Hands-On 17 What is Data Science? 08 What is Machine Learning 09 Supervised and Unsupervised Learning 10 Classification Algorithms 11 Logistic Regression 12 k- nearest neighbors 13 Decision Trees 14 Random Forest 15 Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 7. "Data is the new oil. It’s valuable, but if unrefined it cannot really be used." -Clive Humby Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 8. What is Data Science? Data science is an interdisciplinary field leveraging insights from many fields to extract knowledge from data. 8 https://blog.finxter.com/artificial-intelligence-machine-learning-deep-learning-and-data-science-whats-the-difference/ Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 9. What is Machine Learning? Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 10. Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 11. Classification Algorithms Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 12. Logistic Regression • Extension to the linear regression algorithm • Instead of actual continuous value, we predict the probability of an outcome • Sigmoid Function • It is used when the output is categorical: To predict whether an email is spam (1) or (0) Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation https://towardsdatascience.com/introduction-to-logistic-regression-66248243c148 Get started at: https://ibm.biz/BdfEMZ
- 13. k-nearest neighbors (KNN) • “Birds of a feather flock together.” • Classifies data points based on the points that are most similar to it. • Non-parametric, lazy learning • Classified by majority vote of K nearest neighbors • Assigned to class most common among K neighbors • Examples: Pattern recognition, Recommender systems, image classification Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation https://importq.wordpress.com/2017/11/24/mnist-analysis-using-knn/ Get started at: https://ibm.biz/BdfEMZ
- 14. Decision Trees • Flow-chart like tree structure • uses training data to derive classification rules • Works for both categorical and continuous input and output variables • Non-parametric • Visualization mimics human thinking • Easy to understand and interpret • Prone to over fitting models Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation https://algobeans.com/2016/07/27/decision-trees-tutorial/ Get started at: https://ibm.biz/BdfEMZ
- 15. Random Forest Supervised learning approach • Learning from the input data • Uses the learning to classify new observations • Predicts categorical class labels • Examples: Speech recognition, Face recognition, Car type recognition • Ensemble model that grows multiple tree and classify objects based on the “votes” of all the tree • Fits decision trees on sub-samples of datasets • Uses average to improve accuracy of the model and controls over-fitting • Slow real time prediction • Examples: Fraud prediction, stock market analysis Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation https://towardsdatascience.com/a-quick-and-dirty-guide-to-random-forest-regression- 52ca0af157f8 Get started at: https://ibm.biz/BdfEMZ
- 16. Image source: https://plotly.com/python/ipython-notebook-tutorial/ Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation Get started at: https://ibm.biz/BdfEMZ
- 17. Hands-On Get started at: https://ibm.biz/BdfEMZ • Sign up/Log in to your IBM Cloud Account https://ibm.biz/BdfEMZ Follow along for the hands-on: https://github.com/Anam- Mahmood/watson-studio-learning-path- assets Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation
- 18. Usecase Get started at: https://ibm.biz/BdfEMZ Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation https://www.predicagroup.com/blog/customer-churn-analysis/ • Churn refers to customers no longer using a company’s services or products. • Churn analysis is the evaluation of a company’s customer loss rate in order to reduce it.
- 19. Wrap Up • Machine Learning is a subset of AI techniques which uses statistical methods to enable machines to improve with experience. • Machine Learning has two main types of algorithms: Supervised and Unsupervised. • Jupyter Notebook, as web-based environment for interactive computing. You can run small pieces of code that process your data, and you can immediately view the results of your computation. • Lastly, we have concluded the presentation with an interactive demo for performing data preparation and visualization as well as model training using Jupyter Notebooks. Get started at: https://ibm.biz/BdfEMZ Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation
- 20. Useful Links IBM Developer: https://developer.ibm.com/ Meetup: https://www.meetup.com/IBM-Cloud-MEA/ Learning: https://cognitiveclass.ai/ https://learn.ibm.com/ IBM Watson Studio: https://www.ibm.com/cloud/watson-studio Get started at: https://ibm.biz/BdfEMZ Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation
- 21. Survey https://www.surveygizmo.com/s3/6083679/9e93461cbf3e Get started at: https://ibm.biz/BdfEMZ Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation
- 22. Thank you Sidra Ahmed Developer Advocate, UAE Sidra.Ahmed@ibm.com @acciogibberish Anam Mahmood Developer Advocate, UAE anam.mahmood@ibm.com Get started at: https://ibm.biz/BdfEMY Build and deploy your Machine Learning models effortlessly February 3rd, 2021 / © 2021 IBM Corporation