Abstract
Tanzania's small-scale dairy industry faces similar challenges to those of other developing nations whereby insufficient infrastructure, outdated technology, and low productivity are serious problems for higher milk yield. Tanzania urgently needs to adopt cutting-edge solutions in order to boost dairy performance. With 3500 households' secondary data and 202 households' primary data from 8 villages throughout the Kilimanjaro and Arusha regions, this chapter demonstrates the use of machine learning (ML) techniques to derive homogeneous production clusters and recommendations for more milk yield among dairy farmers. The likelihood for higher milk yield is demonstrated for various clusters with the use of support, confidence, and lift values of association rules analysis. Finally, the production clusters and recommendations are deployed through a mobile application. Recommendations for future improvement are suggested especially on further deployment of learning recommendations and development of a platform-independent mobile solution.
TopGoal two of the Sustainable Development Agenda aims to eradicate hunger, establish food and nutrition security, and advance sustainable agriculture by 2030 (Liverpool-Tasie et al., 2020). To achieve this, it is necessary for small-scale farmers to participate in and gain from the rapid expansion and development of food systems (Liverpool-Tasie et al., 2020). Small-scale livestock producers have an opportunity to improve their livelihoods and become food and nutritionally secure, thanks to the global increase in the demand for and production of animal-source foods, which has been mostly concentrated in Low and Middle Income Countries (LMIC) (four to fivefold increase between 1960 and 2015) (Balehegn et al., 2020).Globally, and especially in LMIC, the demand for animal products is expanding due to population increase, urbanization, and rising affluence (Balehegn et al., 2020).Consumer demand for milk in LMIC will increase by 5.5 million tons, by the year 2050 compared to 2005/2007 (Balehegn et al., 2020). The dairy farming sub-sector has a huge potential to raise people's standards of living (SDG#3) and help end poverty (SDG#1) by promoting better nutrition through milk consumption and revenue from the sale of milk and milk-related goods (Mbilu, 2015; Mzingula, 2019).
Key Terms in this Chapter
Intra-Cluster Similarity: Is a metric for group homogeneity that shows how far apart on average entities are from the cluster's Centre.
Support: Number of instances of an attribute value in the dataset.
Homogenous Groups: A collection of entities that can be objectively distinguished from the other entities and share a common behavior or set of behaviors.
Dairy Production Clusters: A group of dairy farmers that have been identified as using comparable methods, carrying out comparable farm management tasks, and producing comparable results or yields.
Breeding Practices: A selection of animal reproduction techniques that dairy farmers employ for the majority of cattle calving.
Dairy Farming: A dedicated process for raising cattle specifically for the goal of producing milk from hybrid or exotic breeds of cattle.
Confidence: Is a probability of occurrence of an association rule indicating the relationship between the antecedent and consequent.
Supervised Learning: A subfield of machine learning that builds and trains models that can reliably predict and generalize well to new datasets using labelled data.
Recommender system: A data-driven model that has been implemented as an information system to offer recommendations for best practices that could lead to increased yields.
Inter-Cluster Separation: Demonstrates the average distance between the cluster centroids of two or more groups and serves as a gauge of how heterogeneous two groups are.
Unsupervised Learning: A subfield of machine learning that looks for patterns and untapped information in unlabeled datasets.
Peer-to-Peer Learning: A strategy that enables entities with comparable characteristics and patterns of behavior to share information and knowledge with the aim of enhancing performance.
Machine Learning: A collection of methods and tools for creating software programmes that can analyses previous data and make precise predictions about the future.
Small-Scale Farmers: A person who practices agriculture (crop or/and livestock keeping) in a relatively small area of land, normally less than 2 acres.