correction

papers/emily_dorne/main.md

@@ -592,6 +592,6 @@ CyFi is a powerful tool for identifying high and low levels of cyanobacteria, an
 [^footnote-2]: The authors would like to thank Yang Xu, Andrew Wheeler, and Raphael Kimina, the Tick Tick Bloom competition winners whose modeling approaches provided the foundation for CyFi.
 [^footnote-3]: The success of decision tree models in the Tick Tick Bloom competition is consistent with other DrivenData competitions where the task was point-based prediction from satellite imagery (i.e., estimating the [amount of water in snowpack](https://github.com/drivendataorg/snowcast-showdown) and estimating [levels of air pollution](https://github.com/drivendataorg/nasa-airathon/tree/main/pm25)).
 [^footnote-4]: At the equator, a longitude value to 2 decimal degrees is only accurate to around a 1km distance [@decimal_degrees].
-[^footnote-5]: CyFi did not produce cyanobacteria estimates for the remaining 3,258 points in the train set due to a lack of valid satellite data. In order to produce an estimate, there must be at least one satellite image within 30 days prior to the sampling date where cloud pixels account for less than 5% of the pixels in the bounding box around the sampling point.
+[^footnote-5]: CyFi did not use the remaining 3,258 points in the train set due to a lack of valid satellite data. In order to generate features, there must be at least one satellite image within 30 days prior to the sampling date where cloud pixels account for less than 5% of the pixels in the bounding box around the sampling point.
 [^footnote-6]: CyFi did not produce cyanobacteria estimates for the remaining 1,155 points in the test set due to a lack of valid satellite data. In order to produce an estimate, there must be at least one satellite image within 30 days prior to the sampling date where cloud pixels account for less than 5% of the pixels in the bounding box around the sampling point.
 [^footnote-7]: Alternatively, CyFi can be installed with conda (`conda install -c conda-forge cyfi`), which is recommended for M1 Mac users as of July 2024 due to a known issue with the pip installation of LightGBM.