@article{PAR00022498,
  title = {{A}n ensemble machine learning approach for determination of the optimum sampling time for evapotranspiration assessment from high-throughput phenotyping data},
  author = {{K}ar, {S}. and {P}urbey, {V}. {K}. and {S}uradhaniwar, {S}. and {K}orbu, {L}. {B}. and {K}holova, {J}. and {D}urbha, {S}. {S}. and {A}dinarayana, {J}. and {V}adez, {V}incent},
  editor = {},
  language = {{ENG}},
  abstract = {{E}fficient selection of drought-tolerant crops requires identification and high-throughput phenotyping ({HTP}) of the complex functional (especially canopy-conductance) traits that elicit plant responses to continually fluctu-ating environmental conditions. {H}owever, phenotyping of such dynamic physiology-based traits has been immensely challenging especially due to the limited availability of adequate methods that can provide contin-uous measurements of plant-water relations. {T}herefore, gravimetric phenotyping of plants is being increasingly used to allow one-to-one monitoring of plant-water relations and generate continuous evapotranspiration ({ET}) profiles. {T}he gravimetric sensors or load cells can provide {ET} estimates at very high frequencies, e.g. 15-min interval, as chosen by the user. {T}here is however, no study on understanding the optimum frequency or the sampling time at which {ET} needs to be monitored, such that data-redundancy, noise and processing overhead could be reduced. {H}ence, this paper makes a novel attempt in identifying the optimum sampling time for phe-notyping {ET} from load cells time series. {T}he proposed procedure includes an ensemble {M}achine-{L}earning ({ML}) approach for optimizing the sampling time through time series forecasting of {ET} profiles and classification of genotypes using the forecasted {ET} values. {H}igh-frequency load cells data from the {L}easy{S}can, {HTP} platform, {ICRISAT} were used to derive the {ET} profiles at frequencies or scales varying from 15-min to 180-min, followed by {ET} forecasting and classification at each frequency. {F}or both forecasting and classification, an ensemble of three {ML} algorithms i.e. {S}upport {V}ector {M}achines ({SVM}), {A}rtificial {N}eural {N}etwork ({ANN}) and {R}andom {F}orests ({RF}) were leveraged. {C}onsequently, the performance metrics (of both the operations) obtained from the ensemble were used to compute the entropy-based optimum sampling time. {T}he results reveal that 60-min interval {HTP} data could be credibly used for both, forecasting {ET} as well as correctly classifying the genotypes.},
  keywords = {{H}igh throughput phenotyping ; {E}vapotranspiration ; {T}ime series forecasting ; {T}ime series classification ; {E}nsemble machine learning ; {S}ampling time optimization},
  booktitle = {},
  journal = {{C}omputers and {E}lectronics in {A}griculture},
  volume = {182},
  numero = {},
  pages = {105992 [14 p.]},
  ISSN = {0168-1699},
  year = {2021},
  DOI = {10.1016/j.compag.2021.105992},
  URL = {https://www.documentation.ird.fr/hor/{PAR}00022498},
}