Skip to main content
Login | Suomeksi | På svenska | In English

Browsing by Subject "mammal"

Sort by: Order: Results:

  • Pääkkö, Henna (2021)
    Animal personality is described as consistent behavioural variation between individuals over long periods of time. Behaviours often connected to animal personality are such as boldness, aggressiveness, and anxiety. In this thesis, the focus was on the behaviours along the shy-bold axis, containing various degrees of boldness expressing behaviour. The study was conducted by using long-term data from the past 30 years on the banded mongoose (Mungos mungo) population in the Mweya Peninsula in the Queen Elizabeth National Park in Uganda. In particular, I used the data on regular weighing events done within the population. As the weighing is not forced on these individuals, the participation percentage on these events can be used to describe an individual’s boldness. I used the participation percentage as a boldness index (values between 0 and 1) for each individual to describe their position on the shy-bold axis. This index was then used to analyse the differences between sexes, and the fitness effects boldness had on the individuals of this population by using proxies of survival, weight at sexual maturity and lifetime reproductive success (LRS). To determine long-term consistency between individuals, I analysed the repeatability of the boldness index. The repeatability of these values showed we can consider this behaviour as an animal personality. From the fitness analyses, it was concluded that boldness had significant positive effects on the fitness proxies used, proposing that bold individuals have higher fitness in this population. While sex did not affect an individual’s boldness, it had significant interactions with boldness, affecting the strength of fitness effects on individuals in weight at sexual maturity and LRS.
  • Niemi, Mikko Olavi (2020)
    Standard machine learning procedures are based on assumption that training and testing data is sampled independently from identical distributions. Comparative data of traits in biological species breaks this assumption. Data instances are related by ancestry relationships, that is phylogeny. In this study, new machine learning procedures are presented that take into account phylogenetic information when fitting predictive models. Phylogenetic statistics for classification accuracy and error are proposed based on the concept of effective sample size. Versions of perceptron training and KNN classification are built on these metrics. Procedures for regularised PGLS regression, phylogenetic KNN regression, neural network regression and regression trees are presented. Properties of phylogenetic perceptron training and KNN regression are studied with synthetic data. Experiments demonstrate that phylogenetic perceptron training improves robustness when the phylogeny is unbalanced. Regularised PGLS and KNN regression are applied to mammal dental traits and environments to both test the algorithms and gain insights in the relationship of mammal teeth and the environment.