TY - GEN
T1 - Improving farm management optimization
T2 - American Society of Agricultural and Biological Engineers Annual International Meeting 2015
AU - Liao, Wei Ting
AU - Rodriguez, Luis F.
AU - Diesner, Jana
AU - Lin, Tao
PY - 2015
Y1 - 2015
N2 - Farm management optimization models have been developed to support decision making for farming activities including planting, harvesting, handling, drying and storage. However, key decisions for maximizing farming output, timing of planting and other culture tasks, are affected by spatiotemporal changes of weather- related and other spatiotemporal, conditions. Using near real-time information describing actual farming decisions related to the timing of farming events across different regions can make the outputs of farm management optimization models more accurate. This study aims to develop analytical tools to extract near real-time information for identifying actual timing of crop planting and harvesting schedules from publicly available databases. Text mining is applied to identify planting and harvesting schedules for different crops in various regions. The spatiotemporal patterns of farming activities are also analyzed by comparing annual farming schedules across different regions. In order to support the reasoning behind scheduling of culture tasks, semantic networks representing relationships between concepts are constructed for gaining a qualitative understanding of factors contributing towards key agricultural decisions. Network analysis and visualization help characterize semantic networks assisting to highlight important information regarding farming activities. Preliminary results show the planting date of corn was severely delayed in 2013 but ahead of average in 2014 in Illinois. These patterns and schedules are in accordance with National Agricultural Statistics Service (NASS) database. Analyses of semantic networks also indicate rain in April was a main factor for planting delays in 2013. Given data describing actual spatiotemporal farming schedules, as well as information of weather-related conditions, we can gain insights determining changes in the timing of farming events that can potentially be improved inputs and thus constraints for farm management models.
AB - Farm management optimization models have been developed to support decision making for farming activities including planting, harvesting, handling, drying and storage. However, key decisions for maximizing farming output, timing of planting and other culture tasks, are affected by spatiotemporal changes of weather- related and other spatiotemporal, conditions. Using near real-time information describing actual farming decisions related to the timing of farming events across different regions can make the outputs of farm management optimization models more accurate. This study aims to develop analytical tools to extract near real-time information for identifying actual timing of crop planting and harvesting schedules from publicly available databases. Text mining is applied to identify planting and harvesting schedules for different crops in various regions. The spatiotemporal patterns of farming activities are also analyzed by comparing annual farming schedules across different regions. In order to support the reasoning behind scheduling of culture tasks, semantic networks representing relationships between concepts are constructed for gaining a qualitative understanding of factors contributing towards key agricultural decisions. Network analysis and visualization help characterize semantic networks assisting to highlight important information regarding farming activities. Preliminary results show the planting date of corn was severely delayed in 2013 but ahead of average in 2014 in Illinois. These patterns and schedules are in accordance with National Agricultural Statistics Service (NASS) database. Analyses of semantic networks also indicate rain in April was a main factor for planting delays in 2013. Given data describing actual spatiotemporal farming schedules, as well as information of weather-related conditions, we can gain insights determining changes in the timing of farming events that can potentially be improved inputs and thus constraints for farm management models.
KW - Data analysis
KW - Planting and harvesting schedule
KW - Semantic network
KW - Text mining
UR - http://www.scopus.com/inward/record.url?scp=84951968718&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84951968718&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84951968718
T3 - American Society of Agricultural and Biological Engineers Annual International Meeting 2015
SP - 3040
EP - 3049
BT - American Society of Agricultural and Biological Engineers Annual International Meeting 2015
PB - American Society of Agricultural and Biological Engineers
Y2 - 26 July 2015 through 29 July 2015
ER -