Machine Learning applications in Primary Sector

Machine Learning applications in Primary Sector

Posted on Posted in Machine Learning

Detailed below are some of the recent machine learning applications in primary sector of the economy.

Machine Learning Applications in Primary Sector are as follows:

Agriculture and Farming

About 8000 B.C. ago, at the dawn of agriculture, the total number of living human beings on Earth was about 5 million. According to the United Nations, on July 1, 2015 the world’s population was 7.349 billion and is projected to reach 10 billion people in the year 2056 .Studies estimate that global agriculture production needs to increase by 60 -110 %¬† by 2050 to meet the growing demand. Agriculture depends on many volatile factors such as geography, climate, biology, economics and politics. Furthermore, land use involves multiple interactions making it a complex system with environment constraints.

Precision agriculture uses machine learning to feed people by providing a set of tools to gather information and allows farmers to make proper and place specific decisions about agricultural inputs, processing and production.

Scrutinising phenology of plants improves the timing of harvest, yield prediction, pest control and disaster warning. Supervised and unsupervised learning methods use weed discovery based on shape descriptors and early detection of plant diseases based on spectral structures.Supervised self-organizing maps handle existing information from different soil to analyse multi-layer soil facts to figure crop growth and yield.

A recent use of machine learning in geography is Cellular Automaton (CA) to study land use change. Based on the states of neighbouring cells, CA collects cells on a grid of stated shape that changes according to sets of rules and through separate time steps. For example, ThingWorx uses machine learning to automate advanced and predictive analytics for irrigation systems and a network of sensors, to avoid dry soil, the growth of fungus or avoid water regulations in California. The technology learns from data, discovers patterns, builds validated predictive models and sends information to any application.

In Netherlands, Financial Economic Simulation model affects the agriculture sector. However during a review, it was found that the model is static and do not take into account structural changes. Models using Multiple Linear Regression and Neural Networks built and predicted the farm sizes change effectively four-year in the future.

Machine learning methods are applied in aquaculture domain to predict shellfish farm closure, predict algae bloom, set up a new farm and produce habitat maps from seafloor images.

Other uses of Machine Learning applications in Primary Sector such as :

Mining, Quarry and Archaeology

The mining industry uses machine learning for smarter explorations. The data gathered helps in identifying the minerals, the chemical composition and detecting even deeper deposits of minerals. The modelling of mineral deposits and economic impact during the early stages helps mining companies save time and money. It provides safe working conditions for miners and checks the real-time flow of the rocks during blasting operations. For instance, in Iran at Sungun copper mine, statistical analysis techniques predict fly rock distance to help in minimising the related health hazards and other undesirable environmental impacts such as air overpressure and ground vibrations.

Fossil hunting uses neural networks to learn the pattern and make predictions about the data. An example is the use of networks in the Great Divide Basin to distinguish fossil sites through satellite images. The model recognises correctly 79% of the pixels that were known to represent fossil locations and 99% of the pixels flagged by the network held fossils .

Leave a Reply

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