Predicted Cloud Spend / Consumption - an #EnsembleLearning Approach

an @Azure #machinelearning / #deeplearning use case walk through - How to avoid #spuriousCorrelations, understand drivers of cloud spend (consumption) and accurately predicted future spend based on those drivers.

Transcript

6. Data Preparation Data Pre processing Transformations Data collection Data source Identification Initial Dataset creation Feature Engineering Feature extraction Feature transformation Feature selection Model Building Model Architecture Model Stacking Model Validation Picking the right accuracy measures Model Deployment Using Microsoft Azure
7. Identify the right sample based on business requirements Population Selection Noise/Outlier treatment Missing values Data Cleansing Normalization / Log transformation Aggregation and Encoding Data Transformation Constant values/ Zero variance Highly correlated Data Reduction
9. Data Preparation Data Pre processing Transformations Data collection Data source Identification Initial Dataset creation Feature Engineering Feature extraction Feature transformation Feature selection Model Building Model Architecture Model Stacking Model Validation Picking the right accuracy measures Model Deployment Using Microsoft Azure
10. Customer Spend Prior spend Offering Geo Industry Customers with higher historic run rate (spending/rate) are more likely to spend more on Azure Prior Spend Customers subscribed to higher tier offers are likely to spend more on Azure What offer they subscribed to.. Customers from developed countries are likely to have higher spend Customers from tech dominant states/regions are more likely to spend more What country they belong to.. Customer associated with professional services, tech industry are more likely to spend more on Azure Which industry they belong to..
12. NO MAYBE NO YES
14. Data Preparation Data Pre processing Transformations Data collection Data source Identification Initial Dataset creation Feature Engineering Feature extraction Feature transformation Feature selection Model Building Model Architecture Model Stacking Model Validation Picking the right accuracy measures Model Deployment Using Microsoft Azure
15. ENSEMBLE LEARNING
16. Data Preparation Data Pre processing Transformations Data collection Data source Identification Initial Dataset creation Feature Engineering Feature extraction Feature transformation Feature selection Model Building Model Architecture Model Stacking Model Validation Picking the right accuracy measures Model Deployment Using Microsoft Azure
17. Metric Formulae Pro Cons R^2 % the variance in response explained by predictors Can be misleading MAE Outlier resistant No sense of direction RMSE General purpose Metric Outlier sensitive
19. Model Stats: R^2: 0.82 RMSE: 0.5963707 MAE: 0.247357
20. Data Preparation Data Pre processing Transformations Data collection Data source Identification Initial Dataset creation Feature Engineering Feature extraction Feature transformation Feature selection Model Building Model Architecture Model Stacking Model Validation Picking the right accuracy measures Model Deployment Using Microsoft Azure
21. Data collection Data Preparation Feature Engineering Model Building Output Endpoint dashboards
22. Lorem Ipsum Lorem Ipsum 01 04 Data preparation Feature Engineering Deep Learning Benchmark Crowdsourcing Stacking Model Evaluation Model Tuning