Background. Due to an increasing number of credit card defaulters, companies arenow taking greater precautions when approving credit applications. When a customermeets certain requirements, credit card firms typically use their experience todecide whether to grant them a credit card. Additionally, a few machine learningmethods have been applied to support the final decision. Objectives. The aim of this thesis is to compare the accuracy of logistic regressionclassifier, random forest classifier, and support vector classifier with the ensemblebagging classifier for predicting credit card approval. Methods. This thesis follows a method called general experimentation to determinethe most accurate classification technique for predicting credit card approval. Thedataset is taken from Kaggle, which contains information about credit card applications.The selected algorithms are trained with training data and validate themusing validation data then evaluate their performance on the testing data by usingmetrics such as accuracy, precision, recall, F1 score, and ROC curve. Now ensemblelearning bagging technique is applied to combine the predictions of these multiplemodels using majority voting to create an ensemble model. Finally, the performanceof the ensemble model was evaluated on the testing data and compared its accuracyto that of the individual models to identify the most accurate classification techniquefor predicting credit card approval. Results. Among the four selected machine learning algorithms, the random forestclassifier performed better with an accuracy of 88.41% on the testing dataset.The second-best algorithm is the ensemble bagging classifier, with an accuracy of84.78%. Hence, the random forest classifier is the most accurate algorithm for predictingcredit card approval. Conclusions. After evaluating various classifiers, including logistic regression classifier,random forest classifier, support vector classifier, and ensemble bagging, it wasobserved that the random forest classifier outperformed the other models in termsof predicting accuracy. This indicates that the random forest classifier was better atpredicting credit card approval.

No description available.

Syftet med det här arbetet är att undersöka hur väl logistisk regression kan användas för att prediktera vinnaren i en snookermatch. Detta görs med hjälp av statistik över spelarna samt resultat från tidigare matcher och turneringar. En mängd möjliga förklarande variabler presenteras som exempelvis ranking, antal vunna matcher, hemland, typ av turnering, prissumma och omgång (final, semifinal, et cetera). Även tvåvägs-interaktioner mellan variablerna undersöks. Modeller tas fram utifrån hur de presterar i AIC, BIC, residualavvikelse samt Hosmer-Lemeshow-testet. Därefter mäts deras prediktiva förmåga hos ett helt nytt datamaterial med hjälp av noggrannhet, sammanblandningsmatriser och AUC. Resultatet ger flera olika modeller, men den som i slutändan väljs är en modell med bara en förklarande variabel – skillnaden i spelarnas ranking. Modellen hade rätt i sina prediktioner i 60 procent av fallen. Snooker visar sig vara en relativt svårpredikterad sport, jämfört med exempelvis fotboll och hockey, med flera oväntade utfall där den överlägset bättre rankade spelaren förlorade. Modellen är visserligen bättre än vad att godtyckligt gissa vilken spelare som kommer vinna hade presterat, vilket får ses som ett lägsta mått på användbarhet. / This study aims to investigate how well logistic regression can be used to predict the winner in a snooker game. This is done using statistics on the players and results from previous matches and tournaments. A range of possible explanatory variables are presented, such as ranking, number of wins, country, type of tournament, prize money, and round (final, semifinal, et cetera). Two-way interactions between the variables are also examined. Models are developed based on their performance in AIC, BIC, residual deviation, and the Hosmer-Lemeshow test. Then, their predictive ability is measured on an entirely new data set using accuracy, confusion matrices, and AUC. The result produces several different models, but the one ultimately chosen is a model with only one explanatory variable – the difference in the players' rankings. The model was correct in its predictions in 60 per cent of cases. Snooker turns out to be a relatively difficult sport to predict, compared to, for example, football and hockey, with several unexpected outcomes where the significantly better-ranked player lost. The model is at least better than randomly guessing which player would win, which should be seen as the lowest measure of usefulness.

No description available.

No description available.

The study examines the logic and production system of a Swedish sawmill. The aim for the study is to create an understanding and examine how the Lean method value stream mapping can contribute to identifying non-value-adding activities at manufacturing companies in the sawmill industry.  The case study is based on both a quantitative and qualitative approach to answer the study’s research questions. The literature review has been carried out from the theoretical frame of reference. Primary data was collected through observations and semi-structured interviews with employees at the sawmill. The sawmill’s product flow has been mapped through value stream mapping, which was used to identify problematic areas in the production. This was done by comparing lead time, value-creating time, supermarkets, and controls in the production flow. Thematic analysis was used to analyze the results of the data.  The results from the value stream mapping consists of a 23.4-day lead time. Furthermore, the value-creating time that a product is in production amounted to 0.12 % and 79 % of the time non-value-creating activities were in the form of waiting at supermarkets and controls. The result shows that over 99 % of the time was spent on non-value-creating activities during the production.  The study concluded that the value stream mapping calculates value-creating approaches based on the production lead time, which identifies non-value- creating activities. The wait at supermarkets constitutes the largest share of non- value-creating activity in the product flow (79 %). Through semi-structured interviews with employees, non-value-creating activities could be identified based on the seven wastes.

Electric cars are projected to become the vehicles of the future. A major barrier for their expansion is range anxiety stemming from the limited range a typical EV can travel. EV batteries' performance and capacity are affected by many factors. In particular, the decrease in ambient temperature below a certain threshold will adversely affect the battery's efficiency. This research develops deterministic and two-stage stochastic program model for charging stations' optimal location to facilitate the routing decisions of delivery services that use EVs while considering the variability inherent in climate and customer demand. To evaluate the proposed formulation and solution approach's performance, Fargo city in North Dakota is selected as a tested. For the first chapter, we formulated this problem as a mixed-integer linear programming model that captures the realistic charging behavior of the DCFC's in association with the ambient temperature and their subsequent impact on the EV charging station location and routing decisions. Two innovative heuristics are proposed to solve this challenging model in a realistic test setting, namely, the two-phase Tabu Search-modified Clarke and Wright algorithm and the Sweep-based Iterative Greedy Adaptive Large Neighborhood algorithm. The results clearly indicate that the EV DCFC charging station location decisions are highly sensitive to the ambient temperature, the charging time, and the initial state-of-charge. The results provide numerous managerial insights for decision-makers to efficiently design and manage the DCFC EV logistic network for cities that suffer from high-temperature fluctuations. For the second chapter, a novel solution approach based on the progressive hedging algorithm is presented to solve the resulting mathematical model and to provide high-quality solutions within reasonable running times for problems with many scenarios. We observe that the location-routing decisions are susceptible to the EV logistic's underlying climate, signifying that decision-makers of the DCFC EV logistic network for cities that suffer from high-temperature fluctuations would not overlook the effect of climate to design and manage the respective logistic network efficiently.

The purpose of this study was to examine the effect of talker and listener dialect on the validity and reliability of word recognition scores from two sets of Mandarin speech audiometry materials. Four lists of bisyllabic words in Mainland Mandarin and Taiwan Mandarin dialects were administered to 16 participants of each dialect with normal hearing across two test sessions. The performance on materials presented in the native dialect was compared to performance on non-native dialect assessment to determine validity and reliability of test materials. Statistical analysis indicated significant differences between word recognition scores across test sessions, talker and listener dialect, and among lists. However it is unclear if such differences constitute clinically significant differences.

Traditional clinical trials have been inefficient in their methods of dose finding and dose allocation. In this paper a four-parameter logistic equation is used to model the outcome of Bernoulli-response clinical trials. A Bayesian adaptive design is used to fit the logistic equation to the dose-response curve of Phase II and Phase III clinical trials. Because of inherent restrictions in the logistic model, symmetric candidate densities cannot be used, thereby creating asymmetric jumping rules inside the Markov chain Monte Carlo algorithm. An order restricted Metropolis-Hastings algorithm is implemented to account for these limitations. Modeling clinical trials in a Bayesian framework allows the experiment to be adaptive. In this adaptive design batches of subjects are assigned to doses based on the posterior probability of success for each dose, thereby increasing the probability of receiving advantageous doses. Good posterior fitting is demonstrated for typical dose-response curves and the Bayesian design is shown to properly stop drug trials for clinical futility or clinical success. In this paper we demonstrate that an adaptive Bayesian approach to dose-response studies increases both the statistical and medicinal effectiveness of clinical research.

Each year, billions of dollars are spent on missiles and munitions by the United States government. It is therefore vital to have a dependable method to estimate the reliability of these missiles. It is important to take into account the age of the missile, the reliability of different components of the missile, and the impact of different launch phases on missile reliability. Additionally, it is of importance to estimate the missile performance under a variety of test conditions, or modalities. Bayesian logistic regression is utilized to accurately make these estimates. This project presents both previously proposed methods and ways to combine these methods to accurately estimate the reliability of the Cruise Missile.