Metalearning aims to obtain knowledge of the relationship between the mechanism of learning and the concrete contexts in which that mechanisms is applicable. As new mechanisms of learning are continually added to the pool of learning algorithms, the chances of encountering behavior similarity among algorithms are increased. Understanding the relationships among algorithms and the interactions between algorithms and tasks help to narrow down the space of algorithms to search for a given learning task. In addition, this process helps to disclose factors contributing to the similar behavior of different algorithms. We first study general characteristics of learning tasks and their correlation with the performance of algorithms, isolating two metafeatures whose values are fairly distinguishable between easy and hard tasks. We then devise a new metafeature that measures the difficulty of a learning task that is independent of the performance of learning algorithms on it. Building on these preliminary results, we then investigate more formally how we might measure the behavior of algorithms at a ner grained level than a simple dichotomy between easy and hard tasks. We prove that, among all many possible candidates, the Classifi er Output Difference (COD) measure is the only one possessing the properties of a metric necessary for further use in our proposed behavior-based clustering of learning algorithms. Finally, we cluster 21 algorithms based on COD and show the value of the clustering in 1) highlighting interesting behavior similarity among algorithms, which leads us to a thorough comparison of Naive Bayes and Radial Basis Function Network learning, and 2) designing more accurate algorithm selection models, by predicting clusters rather than individual algorithms.

This study examines the safety effects of the improvements made on multi-lane arterials. The improvements were divided into two categories 1) corridor level improvements, and 2) intersection improvements. Empirical Bayes method, which is one of the most accepted approaches for conducting before-after evaluations, has been used to assess the safety effects of the improvement projects. Safety effects are estimated not only in terms of all crashes but also rear-end (most common type) as well as severe crashes (crashes involving incapacitating and/or fatal injuries) and also angle crashes for intersection improvements. The Safety Performance Functions (SPFs) used in this study are negative binomial crash frequency estimation models that use the information on ADT, length of the segments, speed limit, and number of lanes for corridors. And for intersections the explanatory variables used are ADT, number of lanes, speed limit on major road, and number of lanes on the minor road. GENMOD procedure in SAS was used to develop the SPFs. Corridor SPFs are segregated by crash groups (all, rear-end, and severe), length of the segments being evaluated, and land use (urban, suburban and rural). The results of the analysis show that the resulting changes in safety following corridor level improvements vary widely. Although the safety effect of projects involving the same type of improvement varied, the overall effectiveness of each of the corridor level improvements were found to be positive in terms of reduction in crashes of each crash type considered (total, severe, and rear-end) except for resurfacing projects where the total number of crashes slightly increased after the roadway section is resurfaced. Evaluating additional improvements carried out with resurfacing activities showed that all (other than sidewalk improvements for total crashes) of them consistently led to improvements in safety of multilane arterial sections. It leads to the inference that it may be a good idea to take up additional improvements if it is cost effective to do them along with resurfacing. It was also found that the addition of turning lanes (left and/or right) and paving shoulders were two improvements associated with a project�s relative performance in terms of reduction in rear-end crashes. No improvements were found to be associated with a resurfacing project�s relative performance in terms of changes in (i.e., reducing) severe crashes. For intersection improvements also the individual results of each project varied widely. Except for adding turn lane(s) all other improvements showed a positive impact on safety in terms of reducing the number of crashes for all the crash types (total, severe, angle, and rear-end) considered. Indicating that the design guidelines for this work type have to be revisited and safety aspect has to be considered while implementing them. In all it can be concluded that FDOT is doing a good job in selecting the sites for treatment and it is very successful in improving the safety of the sections being treated although the main objective(s) of the treatments are not necessarily safety related.

GenSel is a genetic selection analysis tool used to determine which genetic markers are informational for a given trait. Performing genetic selection related analyses is a time consuming and computationally expensive task. Due to an expected increase in the number of genotyped individuals, analysis times will increase dramatically. Therefore, optimization efforts must be made to keep analysis times reasonable. This thesis focuses on optimizing one of GenSel’s underlying algorithms for heterogeneous computing. The resulting algorithm exposes task-level parallelism and data-level parallelism present but inaccessible in the original algorithm. The heterogeneous computing solution, ReGen, outperforms the optimized CPU implementation achieving a 1.84 times speedup.

Categorization is a complex decision-making process that requires observers to collect information about stimuli using their senses. While research on visual or auditory categorization is extensive, there has been little attention given to tactile categorization. Here we developed a paradigm for studying tactile categorization using 3D-printed objects. Furthermore, we derived a categorization model using Bayesian inference and tested its performance against human participants in our categorization task. This model accurately predicted participant performance in our task but consistently outperformed them, even after extending the learning period for our participants. Through theoretical exploration and simulations, we demonstrated that the presence of sensory measurement noise could account for this performance gap, which we determined was a present factor in participants undergoing our task through a follow-up experiment. Including measurement noise led to a better-fitting model that was able to match the performance of our participants much more closely. Overall, the work in this thesis provides evidence for the efficacy of a tactile categorization experimental paradigm, demonstrates that a Bayesian model is a good fit and predictor for human categorization performance, and underscores the importance of accounting for sensory measurement noise in categorization models. / Dissertation / Doctor of Philosophy (PhD) / The process of categorization is an essential part of our daily life as we encounter various things in the world. Here we explore a model that attempts to explain this process. This model is derived using Bayesian inference and was applied to human behavioural data in a categorization task. We found that the model accounted for most of the performance of our participants but consistently outperformed them. We conducted simulations to explore and demonstrate that this difference is primarily due to the presence of sensory noise in participants. Once we accounted for this noise, we found that our model predicted human performance even more accurately. The work in this thesis demonstrates that a Bayesian Categorization Model which accounts for sensory noise is a good fit and predictor for human performance on categorization tasks.

This study aims to investigate whether or not it is possible to use a machine learning algorithm on eye-tracking data in order to detect early signs of Alzheimer’s disease, which is a type of dementia. Early signs of Alzheimer’s are characterized by mild cognitive impairment. In addition to this, patients with mild cognitive impairment fixate more when reading. The eye-tracking data is gathered in trials, conducted by specialist doctors at a hospital, where 24 patients read a text. Furthermore, the data is pre-processed by extracting different features, such as fixations and difficulty levels of the specific passage in the text. Thenceforth, the features are applied in a naïve Bayes machine learning algorithm, implementing so called leave-one-out cross validation, under two separate conditions; using both fixation features and features related to the difficulty of the text and in addition to this, only using fixation features. Finally, the two conditions achieved the same results - with an accuracy of 64%. Thereby, the conclusion was drawn that even though the amount of data samples (patients) was small, the machine learning algorithm could somewhat predict if a patient was at an early stage of Alzheimer’s disease or not, based on eye-tracking data. Additionally, the implementation is further analyzed through the use of a stakeholder analysis, a SWOT-analysis and from an innovation perspective. / Denna studie syftar till att undersöka huruvida det är möjligt att använda en maskininlärningsalgoritm på eye-tracking data för att upptäcka tidiga tecken på Alzheimer’s sjukdom, vilket är en typ av demens. Tidiga tecken på Alzheimer’s karaktäriseras av mild kognitiv nedsättning. Vidare fixerar patienter med en mild kognitiv nedsättning mer när de läser. Eye-tracking data samlas in i undersökningar genomförda av specialistläkare på ett sjukhus, där 24 patienter läser en text. Därefter förbehandlas datan genom att extrahera olika features, såsom fixeringar och svårighetsnivåer på specifika avsnitt i texten. Efter detta appliceras features i en naïve Bayes maskininlärningsalgoritm som implementerar så kallad leave-one- out cross validation under två separata fall; användande av enbart fixerings features samt användandet av både fixerings features och features för svårighetsgrad för olika avsnitt i texten. Slutligen erhölls samma resultat i båda fallen – med en accuracy på 64%. Därav drogs slutsatsen att även om mängden data (antalet patienter) var liten, kunde maskininlärningsalgoritmen till viss del förutse om en patient var i ett tidigt stadie av Alzheimer’s sjukdom eller inte, baserat på eye-tracking data. Dessutom analyseras implementationen vidare med användning av en intressentanalys, en SWOT-analys och från ett innovationsperspektiv.

No description available.

No description available.

No description available.

No description available.

No description available.