The area of analytics includes what most consider to be the heart of data science, the combination of statistical methods with machine learning, along with optimization, signal processing, network analysis, and other rigorous quantitative methods from a variety of fields.
Although unified by a broad commitment to advanced mathematical methods and algorithms, in reality this is a heterogeneous collection of competing methods and goals. Tensions include inference vs prediction, parametric vs non-parametric (kernel-based) methods, frequentist vs Bayesian statistics, analytic vs algorithmic solutions (including simulations), matrix vs graph-based methods, etc.
- Key tensions: inference vs prediction, analysis vs simulation.
- Common theme: Mathematical models and methods.
- Realm: abstract machinery.
- Keywords: prediction, inference, machine learning, statistics, operations research, AI, experimental design, causality, optimization, knowledge, models, feature engineering, data mining.
- Values: accuracy, precision, validity, truth, convergence, explainability.
Subareas and Courses
- Optimization Theory
- Information Theory
- Linear Models
- Graphical Models
- Data Mining
- Machine Learning
- Deep Learning
- Adversarial Models
- Bayesian Methods in Machine Learning
- Time-Series
- Text Analytics