Project Description: My research interests lie in high dimensional multivariate time series analysis, especially Spatio-temporal data (data observed in both space and time). Time series data (data observed overtime periods) are full of uncertainties and volatilities from various forces. When data is extremely large, and simple, classical modeling techniques often fail. To obtain plausible forecasts and predictions, modeling such data requires careful tuning. My research involves using tensor factorization techniques for dimension reduction and efficient modeling in the presence of missing data. As data become bigger, the classical methods often fail. In my research, we use tensors(multidimensional) instead of matrices (two-dimensional) for model parameters and data characterizations. We develop algorithms (I use R and Julia) that handle computations easier and faster in the presence of missing data. We also propose parameter-free models and algorithms using Bayesian statistics.