This course will introduce you to two modules:
- Time series analysis and forecasting
- Personalization and latent factor modeling
One of the most important topics faced by decision makers in corporate and government agencies is their units’ future performance. This topic is becoming more and more addressable yet challenging with the advent of the big data era. In data analytics, this is time series. Wide variety of tasks that companies face ranging from financial stock market trading to inventory management to sales planning require understanding time trends and issuing time forecasts. While the scale and nature of these time series is diverse, the analysis relies on a single set of principles called “Time Series Analysis and Forecasting”.
In the second half of this course, personalization techniques will be introduced. Traditional business analytics summarizes information from data and generates conclusions. However, with the capability of collecting astronomical amounts of data and conducting analysis with super high-performance computing, we are entering a new era where information is rich enough to achieve personalized analysis and decision support, such as personalized recommendations, marketing, or even medicine. In this course, we will introduce one of the most important personalization techniques, latent factor modeling. The model will be mainly illustrated in the recommender system context, and its generalization to social network and natural language processing will be discussed as well. Eventually, we will circle back to time series models and see how personalized forecasting can be achieved.
Upon completion of this course, you will be able to:
- Analyze and decompose time series data
- Compute forecasts
- Understand how personalization is quantified
- Apply latent factor models to areas demanding heavy personalization