The document discusses various machine learning concepts including exploratory data analysis using huge amounts of data, using decision trees to make predictions on categorical and continuous data, and the benefits of dimensional reduction in dealing with multicollinearity. It also mentions supervised learning algorithms for modeling data with outcome variables, using clustering to represent countries in 2D space, and choosing the best algorithm to fit a given data set.
The document discusses various machine learning concepts including exploratory data analysis using huge amounts of data, using decision trees to make predictions on categorical and continuous data, and the benefits of dimensional reduction in dealing with multicollinearity. It also mentions supervised learning algorithms for modeling data with outcome variables, using clustering to represent countries in 2D space, and choosing the best algorithm to fit a given data set.
The document discusses various machine learning concepts including exploratory data analysis using huge amounts of data, using decision trees to make predictions on categorical and continuous data, and the benefits of dimensional reduction in dealing with multicollinearity. It also mentions supervised learning algorithms for modeling data with outcome variables, using clustering to represent countries in 2D space, and choosing the best algorithm to fit a given data set.
Huge amount of date with many dimensions- Exploratory data analysis
2. Decision tree can be used for predictions of both categorial and continuous data- true 3. Benefits of dimensional reduction- it deals with multi collinearity 4. Broom on how to sweep the floor- Supervised learning 5. Supervised Algorithm are used for-modelling data that does not come with outcome variables 6. Country is represented as a 2d space-Clustering 7. Boss gives you data set-Choose an algorithm that is the best to fit the data 8. Fraud Detection Module-Tanom 9. ransition classification 10. Random forest and Black box-the vectors of data is so dense they resemble black boxes 11. Programming language used for machine learning-HTML 12. Data with high kurtosis- Heavy tails 13. Creating data with existing data points is – Feature Creation 14. HIV Detection-all in days work Nothing to extremely happy about it/ of course happy about it 15. Dependent Variable-the output or outcome whose variation is studied 16. Confusion matrix-type2 Is false positive, type 1 is false negative 17. Univariate analysis- an analysis that involves single variable 18. Project manager skeptical about allotting time-Dimensionality reduction removes redundant features but takes care of multilinearity 19. Writing algorithm, -the value of b, if a is 7 approximately 17 20. After conducting machine learning algorithms to the data set-Correlation 21. Machine learning model to boss-Bias is error due to erroneous or overly complex assumptions 22. Transforming of continuous data to categorical order- equal frequency 23. Example of continuous variable – time 24. Anomaly detection can be supervised or un supervised-false/ true 25. Linear regression are not considered for machine learning-False 26. Common goal of machine learning-writing macros to automate simple computer tasks 27. Geographical data – Unsupervised learning 28. Outlier creates a significant association- report any significance from your analysis 29. Mean imputation is the best method- true 30. Managing platform database -Histogram 31. Trainee asked advice – Neural networks 32. Client gave you dataset for your machine learning-place a value perhaps based on mean, or other metrics that are suitable for domain 33. Cognitive system- C 34. When one predictor variable in multiple regression model can be linearity- Multicollinearity 35. New senior data scientist- data cleaning 36. Fresh hires for artificial intelligence- reinforcement learning 37. Graphical representation-(300,7) &(600,10) 38. Current prjt manager-Dimensionality reduction quickens the process of performing computations because less dimensions lead to less space