Skip to main content
ARS Home » Plains Area » Las Cruces, New Mexico » Range Management Research » Research » Publications at this Location » Publication #354318
Title: Using Machine Learning to Increase Research Efficiency: A New Approach in Environmental Sciences

Author
item RAMIREZ, GEOVANY - New Mexico State University
item Peters, Debra
item JOPPA, LUCAS - Microsoft

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 6/21/2018
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: Data collection has evolved from tedious in-person fieldwork to automatic data gathering from multiple sensor remotely. Scientist in environmental sciences have not fully exploited this data deluge, including legacy and new data, because the traditional scientific method is focused on small, high quality datasets instead of more complex and larger datasets. We present a system that helps with the implementation of the new scientific approach based on a knowledge-driven, open access system that learns and becomes more efficient and easier to use as data streams increase in variety and size. Our Knowledge, learning, and analysis system (KLAS) implemented a recommendation system based on multiple users' behavior using machine learning to serve as a guide during the experimentation process. The learning mechanism should be able to improve the accuracy and quality of recommendation as more users interact with KLAS. We implemented tools to help the scientific community to reuse data, methods, and models. Another feature of KLAS is the ability to improve efficiency of field-collected data. For instance, we have interconnected KLAS with an automatic system for data harvesting and data QA/QC from a network of meteorological stations. Users can easily perform experiments with data collected by sensors spatially distributed in the field, and user decisions can be guided by past user experiences. In addition to simplifying the research process, we see KLAS as a learning tool for students where they can learn from experienced users.