Joseph Gitlin L&S Math & Physical Sciences
A New General Estimator and its Foundations
Estimators give us an idea of the true value of our estimand, which represents a phenomenon that we observe in the real world. The concern with estimators is whether or not they follow certain properties, which is checked by underlying math such as Analysis and Probability Theory, and how accurately and consistently they capture the real value. With more data, it becomes more apparent which estimators are accurate with low variance; it’s easier to see how to improve these estimators given more data used for fitting them to the observations. However, when fitted data is changed and the estimator is applied to a new data set, the relationship given by the estimator tends to be less accurate and perform worse. This research seeks the possibility and discovery of a new general estimator that has low error and variance when applied to a new data set, even a data set that follows a completely different distribution. A new general estimator could potentially be an ubiquitous analysis tool that will require minimal computational power and little to no need of auspicious fitted data when estimating a certain phenomenon. It can provide more accurate estimation in the general analysis of any field.