How our models can be used in various sustainability challenges
In this section of our research, our team created a machine learning model that can help predict what type of crop to plant in certain conditions. We then created a model competition notebook that compared three popular algorithms, eventually determining the one that outputs the best results for our intended problem. Included are also discoveries our team has made on how our model helps in other ways such as socioeconomic or conservation situations.
Another feature of our website is making crop recommendations based on climate data, like humidity, moisture levels, temperature, and levels of different substances in the soli (K, N, P) that the landowners can provide. We used several classifier models applied to the same data set and compared those to choose the best. Models implemented were KNN, Decision Tree, Random Forest, Linear SVC, Gradient Boosting, Poly SVC, Grid Search, and RBF SVC. After comparing the models, we choose to use Gradient Boosting, which has a 99.27% accuracy. By facilitating agriculturalists with a platform where they can receive information regarding the species of plants that could thrive in their particular conditions, we maximize production and help against the decline in crop yield. Furthermore, our solution also proposes a game that can educate the population about the effects of climate change on agriculture. Due to the pollution caused by fossil fuel combustion, carbon dioxide levels are rising, reducing the concentration of nutrients in crops. Our website tackles the problem of minimizing the damage caused by this phenomenon by providing information about efficiently using our already existing resources.
Climate change is one of the biggest and most important issues in our daily lives. As the global average temperature has increased over the past years, there have been large impacts that directly affect our health, environment, and economy. Not only have there been changes in rainfall in several places, which in consequence results in floods, droughts, or intense rain, but also more frequent heatwaves. Climate change can also directly affect our health by worsening air and water quality, which can potentially increase the spread of diseases. Animals are also directly affected as the changing ecosystem may influence the timing of major events in their lifecycle, such as migration or reproduction. These countless consequences due to climate change all stem from an increased amount of carbon dioxide in our atmosphere. Plants, though can help reduce the spread of climate change. Through photosynthesis, plants suck carbon dioxide from the atmosphere and convert it into their own food; and because of this, plants are considered to be some of the most important carbon sinks. Since carbon dioxide stays in the air decades longer than other toxic greenhouse gasses, efforts to reduce it are extremely critical to mitigating climate change. Thus as plants take up carbon dioxide from the air, they not only help improve air quality but also can aid in slowing down climate change. And as more and more human activities are causing carbon dioxide to be emitted into the atmosphere, scientists have found that in response, plants have been photosynthesizing more. While global carbon dioxide concentrations grew about 17% between 1982 to 2020, there was also a 12% increase in global photosynthesis. Though 12% is a very large increase in photosynthesis, it is still nowhere close to removing the extensive amounts of carbon dioxide that is emitted into the atmosphere daily. Regardless, every plant planted can help reduce climate change. Our machine learning algorithm that gathers information about specific crops will lead to more healthy and thriving plants, which in turn can aid in reducing climate change.
Climate change is reducing crop production due to many factors including crop diseases. This then decreases food availability to both humans as well as other animals. The change in temperature, precipitation patterns, and rate of extreme weather events as a result of climate change all impact crop yields around the world. With the growing population, the demand for food is rising. As a result, climate change as a whole is expected to increase rates of food insecurity. Food insecurity is the insubstantial or uncertain access to sufficient food. In 2019, two out of the almost eight billion people on earth experienced food insecurity. We used computer vision to detect whether crops are healthy or infected by disease. We also compared the accuracy of models being able to classify what the best type of crop to grow is based on climate data. The Gradient Boosting model was the most accurate. Gradient Boosting is an ensemble technique where predictors learn from the errors of the predictors before them. Farmers can potentially use these models to improve crop yield by choosing the most successful crops to grow in the first place and dealing with diseased crops as soon as it occurs. By identifying which plants have diseases, further steps can be taken to prevent diseases from occuring in the first place. In turn, this can improve crop yields and decrease the rate of food insecurity.
Climate Change. This is a phrase which is underestimated by few and plays a huge role in others’ lives. It is buzzing across nations as a UN SDG, as people hassle to find the best way to reduce its detrimental effects on society. In short, it can make living a healthy life a luxury with rising sea levels, changing weather patterns, and food shortage. Imagine a life without sufficient food and nutrients to keep nearly 8 billion people alive. Seems almost impossible. With climate change forcing the food supply and population to grow in opposite directions, we devised a solution of our own to battle against climate change. We will use AI models and Computer Vision to detect crop diseases. With a simple solution to a large-scale problem, we tested multiple models like Random Forest Classifier and Gradient Boosting to evaluate the pros and cons for the best one. The Gradient Boosting model we created had an accuracy of over over 95% that is available on our website that helps make recommendations to increase sustainability of crop production today. Our interactive features extend beyond this as we created a video game to increase awareness on the numerous problems that arose as a result of climate change, including the lack of sufficient food. Overall, our idea to detect crop diseases and create awareness on the gravity of the problem, helps reduce the uncertainty people bear regarding food availability, stimulates people to take action, and alleviates farmers and millions of people across the world struggling to find their daily meals.