Wazihub Soil Moisture Prediction Challenge
$8,000 USD
Predict soil humidity using sensor data from low-cost DIY Internet of Things in Senegal
723 data scientists enrolled, 96 on the leaderboard
29 July 2019—21 October 2019
NO FEATURES TO INFER
published 10 Sep 2019, 11:08

For field1 (Maize) there are no features to make inference with , they are all NULL , even in the context data the 4 days to predict are without features. Is it a problem with files or what???!!!

dude there is something very weird with the data, the test data is even weirder, I don't know if that's something to solve ourselve or shoud we just ignore this entirely. beside what do they mean with future data which we have no right to use.?

this is very confusing.

did you ever get to the bottom of this? I am finding it a bit counter intuitive if we are training models using the atmospheric data given and then feeding the test data to the model without any of the conditions to make the predictions from?

edited 1 minute later

Aha, I get it now , I was confused with missing values in test data and how to predict them , it seems that time series forcasting is different from machine learning prediction and i have never play with time series data before . you can check this link: https://blogs.oracle.com/datascience/7-ways-time-series-forecasting-differs-from-machine-learning