This thesis investigates imbalanced Swedish text financial datasets, specifically receipt classification using machine learning models. The study explores the effectiveness of under-sampling and over-sampling methods for Naive Bayes algorithms, collaborating with Fortnox for a controlled experiment. Evaluation metrics compare balancing methods regarding the accuracy, Matthews's correlation coefficient (MCC) , F1 score, precision, and recall. Findings contribute to Swedish text classification, providing insights into balancing methods. The thesis report examines balancing methods and parameter tuning on machine learning models for imbalanced datasets. Multinomial Naive Bayes (MultiNB) algorithms in Natural language processing (NLP) are studied, with potential application in image classification for assessing industrial thin component deformation. Experiments show balancing methods significantly affect MCC and recall, with a recall-MCC-accuracy tradeoff. Smaller alpha values generally improve accuracy.  Synthetic Minority Oversampling Technique  (SMOTE) and Tomek's algorithm for removing links developed in 1976 by Ivan Tomek. First Tomek, then SMOTE (TomekSMOTE)  yield promising accuracy improvements. Due to time constraints, Over-sampling using SMOTE and cleaning using Tomek links. First SMOTE, then Tomek (SMOTETomek) training is incomplete. This thesis report finds the best MCC is achieved when $\alpha$ is 0.01 on imbalanced datasets.

OBJECTIVE: Metamemory is integral for strategizing about memory intentions. This study investigated the prospective memory (PM) deficit in Parkinson's disease (PD) from a metamemory viewpoint, with the aim of examining whether metamemory deficits might contribute to PM deficits in PD. METHOD: Sixteen patients with PD and 16 healthy older adult controls completed a time-based PM task (initiating a key press at two specified times during an ongoing task), and an event-based PM task (initiating a key press in response to animal words during an ongoing task). To measure metamemory participants were asked to predict and postdict their memory performance before and after completing the tasks, as well as complete a self-report questionnaire regarding their everyday memory function. RESULTS: The PD group had no impairment, relative to controls, on the event-based task, but had prospective (initiating the key press) and retrospective (recalling the instructions) impairments on the time-based task. The PD group also had metamemory impairments on the time-based task; they were inaccurate at predicting their performance before doing the task but, became accurate when making postdictions. This suggests impaired metamemory knowledge but preserved metamemory monitoring. There were no group differences regarding PD patients' self-reported PM performance on the questionnaire. CONCLUSIONS: These results reinforce previous findings that PM impairments in PD are dependent on task type. Several accounts of PM failures in time-based tasks are presented, in particular, ways in which mnemonic and metacognitive deficits may contribute to the difficulties observed on the time-based task.

The survival rate for a suddent cardiac arrest (SCA) is incredibly low, with less than one in ten surviving; most SCAs occur outside of a hospital setting. There is a need to develop an effective and efficient system that can sense, communicate and remediate potential SCA situations on a near real-time basis. This research presents a novel Zeolite-PDMS-based optically unobtrusive flexible dry electrodes for biosignal acquisition from various subjects while at rest and in motion. Two zeolite crystals (4A and 13X) are used to fabricate the electrodes. Three different sizes and two different filler concentrations are compared to identify the better performing electrode suited for electrocardiogram (ECG) data acquisition. A low-power, low-noise amplifier with chopper modulation is designed and implemented using the standard 180nm CMOS process. A commercial off-the-shelf (COTS) based wireless system is designed for transmitting ECG signals. Further, this dissertation provides a framework for Machine Learning Classification algorithms on large, open-source Arrhythmia and SCA datasets. Supervised models with features as the input data and deep learning models with raw ECG as input are compared using different methods. The machine learning tool classifies the datasets within a few minutes, saving time and effort for the physicians. The experimental results show promising progress towards advancing the development of a wireless ECG recording system combined with efficient machine learning models that can positively impact SCA outcomes.

In the last decades, computational models have become useful tools for studying biological neural networks. These models are typically constrained by either behavioural data from neuropsychological studies or by biological data from neuroscience. One model of the latter kind is the Bayesian Confidence Propagating Neural Network (BCPNN) - an attractor network with a Bayesian learning rule which has been proposed as a model for various types of memory. In this thesis, I have further studied the potential of the BCPNN in short-term sequential memory. More specifically, I have investigated if the network can be used to qualitatively replicate behaviours of immediate verbal serial recall, and thereby offer insight into the network-level mechanisms which give rise to these behaviours. The simulations showed that the model was able to reproduce various benchmark effects such as the word length and irrelevant speech effects. It could also simulate the bow shaped positional accuracy curve as well as some backward recall if the to-be recalled sequence was short enough. Finally, the model showed some ability to handle sequences with repeated patterns. However, the current model architecture was not sufficient for simulating the effects of rhythm such as temporally grouping the inputs or stressing a specific element in the sequence. Overall, even though the model is not complete, it showed promising results as a tool for investigating biological memory and it could explain various benchmark behaviours in immediate serial recall through neuroscientifically inspired learning rules and architecture. / Under de senaste årtionden har datorsimulationer blivit ett allt mer populärt verktyg för att undersöka biologiska neurala nätverk. Dessa modeller är vanligtvis inspirerade av antingen beteendedata från neuropsykologiska studier eller av biologisk data från neurovetenskapen. En modell av den senare typen är ett Bayesian Confidence Propagating Neural Network (BCPNN) - ett autoassociativt nätverk med en Bayesiansk inlärningsregel, vilket tidigare har använts för att modellera flera typer av minne. I det här examensarbetet har jag vidare undersökt om nätverket kan användas som en modell för sekventiellt korttidsminne genom att undersöka dess förmåga att replikera beteenden inom verbalt sekventiellt korttidsminne. Experimenten visade att modellen kunde simulera ett flertal viktiga nyckeleffekter såsom the word length effect och the irrelevant speech effect. Däröver kunde modellen även simulera den bågformade kurvan som beskriver andelen lyckade repetitioner som en funktion av position, och den kunde dessutom repetera korta sekvenser baklänges. Modellen visade också på viss förmåga att hantera sekvenser där ett element återkom senare i sekvensen. Den nuvarande modellen var däremot inte tillräcklig för att simulera effekterna som tillkommer av rytm, såsom temporär gruppering eller en betoning på specifika element i sekvensen. I sin helhet ser modellen däremot lovande ut, även om den inte är fullständig i sin nuvarande form, då den kunde simulera ett flertal viktiga nyckeleffekter och förklara dessa med hjälp av neurovetenskapligt inspirerade inlärningsregler.

No description available.

No description available.

No description available.

No description available.

The issue of food scandals originating from unsafe food has in recent years caught the attention of the public. Consequently, the number of food recalls has increased, and food retailers and food producers are getting questioned regarding their ability to provide consumers with safe food products. Thus, because the primary responsibility to deal with this problem is the food business operators who have real control over the products within the Food Supply Chain (FSC). Even though food safety has received more interest from academic researchers and practitioners in the latter years, we could identify that previous literature still lacks research on food recalls. In addition, prior research interprets food recalls as generalizable within the FSC and does not have its various product flows in mind. In order to fill the research gap, this study seeks to gain a deeper understanding of food recalls in the context of the different product flows within the FSC from a retailer perspective. Therefore, the purpose of this study was to investigate how the occurrence and the consequences of food recall differ across various product flows within the FSC by focusing on four elements. The elements include the challenges within the FSC to ensure food quality, the reasons behind food recall, the types of food recall, and the consequences they infer. The current study had finally answered the following research question: “How do the occurrence and the consequences of food recalls differ across various product flows within the FSC?” In order to answer the research question and address the purpose of the thesis, a qualitative study was conducted by focusing on Swedish food retailers and conducting interviews with responsible Purchasers within each of the product flows, and additionally, Quality Assurance Managers. The sample of the semi-structured interviews was selected based on the position and experience of food recalls within the retailers. This to assure that the interviewees possessed sufficient insights regarding the studied field to fulfill the purpose of the study and answer its research question. From the gathered data, several themes were derived through the thematic analysis, and the analysis and discussion regarding the elements led us to a conclusion. The result confirmed that the various product flows, and in some cases, even specific products within the flows, need to be taken into consideration when the occurrence and the consequences of food recalls are addressed within the FSC. This is because the different product flows include different products with different characteristics making them more or less complex to handle along the chain and when carrying out food recalls. Even though previous theories assume a general approach, the findings still confirmed the theories regarding the challenges within the FSC to ensure food quality, the reasons behind recalls, the types of recalls, and the consequences the recalls can infer. However, the main differences found between our study and prior research regarding the FSC were that supply chain-related reasons are more commonly related to food reclaims and that environmental consequences are considered an additional consequence of food recalls.

Adolescents are especially vulnerable to image-based messages, which place them directly in line with the lifestyle messages presented in numerous alcohol advertisements. Research has shown that teenagers encounter alcohol advertisements from a variety of sources, prompting concern over the consequences of exposure and the risks associated with teen drinking. Warning messages contained within these alcohol advertisements are often vague in their wording and targeted to promote responsible drinking among adults, making them wholly inadequate for teen deterrence. This thesis explores the problems with current alcohol responsibility messages when dealing with teens, and proposes the development of new, modified messages, that are more direct and consequence-based. The Elaboration Likelihood Model of persuasion was used to describe teen processing of alcohol warning messages, and eye-tracking technology was used to determine how this processing can be examined visually. The purpose of this research was to aid in the development of new messages that have been modified by type size, type contrast, and message content (a 2x2x2 factorial design). Results indicated that modifying message content to include direct language and a consequence enhances visibility and attention to alcohol responsibility messages. Larger type and higher contrast can be applied to the message to further enhance visibility, but the change in attention for these factors was shown to be minimal and nonsignificant. Overall, the greatest power lies in changing the message itself, irrespective of all other variables.