The heterogeneity of smallholder dairy production systems complicates service provision, information sharing, and dissemination of new technologies, especially those needed to maximize productivity and profitability. In order to obtain homogenous groups within which interventions can be made, it is necessary to define clusters of farmers who undertake similar management activities. This paper explores robustness of production cluster definition using various unsupervised learning algorithms to assess the best approach to define clusters. Data were collected from 8179 smallholder dairy farms in Ethiopia and Tanzania. From a total of 500 variables, selection of the 35 variables used in defining production clusters and household membership to these clusters was determined by Principal Component Analysis and domain expert knowledge. Three clustering algorithms, K-means, fuzzy, and Self-Organizing Maps (SOM), were compared in terms of their grouping consistency and prediction accuracy. The model with the least household reallocation between clusters for training and testing data was deemed the most robust. Prediction accuracy was obtained by fitting a model with fixed effects model including production clusters on milk yield, sales, and choice of breeding method. Results indicated that, for the Ethiopian dataset, clusters derived from the fuzzy algorithm had the highest predictive power (77% for milk yield and 48% for milk sales), while for the Tanzania data, clusters derived from Self-Organizing Maps were the best performing. The average cluster membership reallocation was 15%, 12%, and 34% for K-means, SOM, and fuzzy, respectively, for households in Ethiopia. Based on the divergent performance of the various algorithms evaluated, it is evident that, despite similar information being available for the study populations, the uniqueness of the data from each country provided an over-riding influence on cluster robustness and prediction accuracy. The results obtained in this study demonstrate the difficulty of generalizing model application and use across countries and production systems, despite seemingly similar information being collected.

Artificial insemination (AI) and selective bull mating are considered as robust methods for dairy cattle breeding. Globally, these methods have been used to enhance productivity and realize rapid genetic gains. However, these technologies have had low adoption rates in sub-Saharan Africa (SSA). Even though available evidence suggests that this is due to various infrastructural and technical challenges. There is limited information about what drives this low uptake of AI from a farmer's perspective. Therefore, the main objective of this study was to determine and characterize factors that influence the choice by smallholder farmers between bull service and AI for dairy cow breeding. Further, the relationships between the breeding choices and the bio-physical elements of dairy farming, mainly, farmer characteristics, household income levels, farm management practices, and institutional support structures, were investigated. Data were collected through face-to-face interviews from a total of 16,308 small-scale dairy farmers in Ethiopia (n = 4679), Kenya (n = 5278), Tanzania (n = 3500), and Uganda (n = 2851). The questionnaire was coded in an electronic form using Open Data Kit (ODK) platform to allow for real-time data entry and management. Descriptive statistics, chi-square test, and a t-test were used to evaluate the independent and dependent variables, while logistic regression and factor analysis were used to identify factors that influenced farmers' breeding decisions. Results showed that there was a significant difference in animal husbandry practices between farmers who used artificial insemination (AI) and those who practiced bull mating. The majority of farmers who used AI kept records, purchased more animal feeds, had more labor by hiring workers whose average wages were higher than those of bull service farmers. However, farmers who used AI pay more for services such as water access and breeding while their service providers had to cover long distances compared to farmers who used bulls. This indicates limited access to services and service providers for AI farmers. The ratio of AI to bull service users was even for Ethiopia and Kenya, while in Uganda and Tanzania, more farmers preferred bull service to AI. It was established that the factors that influence farmers' breeding decision were not the same across the region. Factors such as farmer's experience in dairy farming, influence of the neighbor, farmer's ability to keep records, and management practices such as water provision and availability of feeds had a significant association (p < 0.001) with AI adoption among dairy farmers. In contrast, large herd and large land size negatively influenced AI adoption. Institutional settings including cost of AI service and the distance covered by the service provider negatively affected (p < 0.001) the choice of AI as a breeding option. There was a high probability of continued use of a specific breeding method when there was a previous conception success with that same method. Based on the results obtained, we...

Characterization of smallholder farmers has been conducted in various researches by using machine learning algorithms, participatory and expert-based methods. All approaches used end up with the development of some subgroups known as farm typologies. The main purpose of this paper is to highlight the main approaches used to characterize smallholder farmers, presenting the pros and cons of the approaches. By understanding the nature and key advantages of the reviewed approaches, the paper recommends a hybrid approach towards having predictive farm typologies. Search of relevant research articles published between 2007 and 2018 was done on ScienceDirect and Google Scholar. By using a generated search query, 20 research articles related to characterization of smallholder farmers were retained. Cluster-based algorithms appeared to be the mostly used in characterizing smallholder farmers. However, being highly unpredictable and inconsistent, use of clustering methods calls in for a discussion on how well the developed farm typologies can be used to predict future trends of the farmers. A thorough discussion is presented and recommends use of supervised models to validate unsupervised models. In order to achieve predictive farm typologies, three stages in characterization are recommended as tested in smallholder dairy farmers datasets: (a) develop farm types from a comparative analysis of more than two unsupervised learning algorithms by using training models, (b) assess the training models’ robustness in predicting farm types for a testing dataset, and (c) assess the predictive power of the developed farm types from each algorithm by predicting the trend of several response variables.

No abstract

Peer-to-peer learning paradigm is seldom used in studying how farmers can increase yield. In this article, agent-based modelling has been applied to study the chances of dairy farmers increasing annual milk yield by learning better farming strategies from each other. Two sets of strategies were considered; in one set ( S), each farmer agent would possess a number of farming strategies based on their knowledge, and in a second set [Formula: see text], farmer agents would possess farming strategies that they have adopted from their peers. Regression models were used to determine litres of milk that could be produced whenever new strategies were applied. By using data from Ethiopia and Tanzania, 28 and 25 determinants for increase in milk yield were fitted for the two countries, respectively. There was a significant increase in average milk yield as the farmer agents interacted and updated their [Formula: see text]– from baseline data, average milk yield of 12.7 ± 4.89 and 13.62 ± 4.47 to simulated milk yield average of 17.57 ± 0.72 and 20.34 ± 1.16 for Tanzania and Ethiopia, respectively. The peer-to-peer learning approach details an inexpensive method manageable by the farmers themselves. Its implementation could range from physical farmer groups to online interactions.