The ability to obtain accurate food security metrics in developing areas where relevant data can be sparse is critically important for policy makers tasked with implementing food aid programs . As a complement to existing techniques in crop yield prediction, this work develops neural network models for predicting the sentiment of Twitter feeds from farming communities . We find that direct learning from small, relevant datasets outperforms transfer learning from large, fully-labeled datasets, and convolutional neural networks broadly outperform recurrent neural networks on Twitter sentiment classification . These models perform substantially less well on ternary sentiment problems characteristic of practical settings, and that these models perform significantly less well in the . practical settings than on binary problems often found in the literature . We hope these models would ultimately serve as a useful crop yield predictors. of the food security predictor. We also find that these . models perform less well to the ternarian sentiment problems characterized by practical settingsĀ and that they perform substantially .

Links: PDF - Abstract

Code :

None

Keywords : models - sentiment - perform - food - problems -

Leave a Reply

Your email address will not be published. Required fields are marked *