Method for estimating a spatial distributions function from a number of measurement data.
We often must try to estimate and predict an overview image of a measurable quantity in a space based on measurement data at discrete points. In Earth Science, it is sometimes difficult to take the actions necessary to measure in the field and, as a result, obtaining data becomes cumbersome. As a result, we are often forced to work with insufficient data. On the other hand, because of the progress in modern technology, we are now able to obtain large quantities of data. In this case, we must be systematic when processing such large quantities of data in order to obtain objective and accurate estimations, not ones that are biased. This paper describes universal approaches to estimating and predicting the distribution function of measurable quantity in space from a finite number of measurement data at discrete points with measurement errors. I assume the nature of “well behavior in space” for the distribution function, which means differentiability in the space. A statistical approach for the prediction is employed by choosing the best solution of the probability distribution functions for the quantities to be estimated.
| Main Author: | |
|---|---|
| Format: | Journal Contribution biblioteca |
| Language: | Japanese |
| Published: |
2019
|
| Online Access: | http://hdl.handle.net/1834/15466 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|