Awesome

Yield-Prediction-DNN

This repository contains my code for the "Crop Yield Prediction Using Deep Neural Networks" paper authered by Saeed Khaki and Lizhi Wang. The network is a deep feedforward neural network which uses the state-of-the-art deep learning techniques such as residual learning, batch normalization, dropout, L1 and L2 regularization.

Getting Started

Please install the following packages in Python3:

• numpy

• tensorflow

• matplotlib

Dimension of Input Data

• Genotype , soil, and weather data were used in the paper. You should load your data as train and test, then run the model.

• The genotype data were coded in -1, 0, 1 values, respectively representing aa, aA, and AA alleles. The genotype data had dimension n-by-p where n and p denote the number of obseration and genetic markers.

• The environment data (weather and soil) had dimension n-by-k, where n and k denote the number of obseration and enviromental components.

• Each observation is fed to the network as 1d vectors (dimension p+k) .

Data Availability Statement

The data analyzed in this study was provided by Syngenta for 2018 Syngenta Crop Challenge. We accessed the data through annual Syngenta Crop Challenge. During the challenge, September 2017 to January 2018, the data was open to the public. Researchers who wish to access the data may do so by contacting Syngenta directly. We are not allowed to share the data due to non-disclosure agreement, sorry.

Notice

We have recently published a new paper titled <a href="https://arxiv.org/abs/1911.09045" target="_blank">"A CNN-RNN Framework for Crop Yield Prediction"</a> published in <a href="https://www.frontiersin.org/articles/10.3389/fpls.2019.01750/abstract" target="_blank"> Frontiers in Plant Science Journal</a>. This paper predicts corn and soybean yields based on weather, soil and management practices data. Researchers can use the data from this paper using following <a href="https://github.com/saeedkhaki92/CNN-RNN-Yield-Prediction" target="_blank"> link</a>. We spend a lot of time gathering and cleaning the data from different publicly available sources. Please cite our papers if you use our data or codes. Thanks.