Machine learning is becoming a part of everyday life and has an indisputable impact across large array of industries. In the financial industry, this impact lies particularly in predictive modelling. The goal of this thesis is to describe the basic principles of artificial intelligence and its subset, machine learning. The most widely used machine learning techniques are outlined both in a theoretical and a practical way. As a result, four models were assembled within the thesis. Results and limitations of each model were discussed and these models were also mutually compared based on their individual per- formance. The evaluation was executed on a real world dataset, provided by Home Credit company. Final performance of machine learning methods, measured by the KS and GINI metrics, was either very comparable or even worse than the performance of a traditional logistic regression. Still, the problem may lie in an insu cient dataset, in the improper data prepara- tion, or in inappropriately used algorithms, not necessarily in the models themselves.

No description available.

Thesis submitted in fulfillment of the requirements for the degree of Master of Management in Finance & Investment In the Faculty of Commerce and Law Management Wits Business School at the University of the Witwatersrand, 2015 / Risk based pricing has been a topic of discussion since the 2008 financial crisis as a result of the on-selling of packaged sub-prime assets. This paper will highlight the importance of correctly assessing risk within the framework of consumer credit provision. We will begin with a brief overview of the South African unsecured lending market, look into the definition of risk based pricing and the impact it has had in the market and conclude the paper by using a model by Robert Phillips to calculate the interest rate to be offered to a customer. / AC2016

This study investigated the self-efficacy (also often referred to as self-confidence) of principals as determined by school administrator certification credentials and teaching endorsements at low performing middle schools in Mississippi. In educational literature, the term “self-confidence” is often referred to under the nomenclature of self-efficacy. In the context of an educational environment, self-efficacy pertains to a principal’s capability to organize and execute courses of action required in leading and managing a school. Successful school management requires a leader who is task oriented, consistently stays focused, employs effective strategies, and utilizes managerial skills. The investigation focused on the self-efficacy, as determined by credentials and endorsements, of the principals charged with leading and managing the 24 Mississippi middle schools that received Mississippi Department of Education accountability ratings of “D” or “F” in relation to student academic performance. The overall research question that guided the investigation asked: Did the self-efficacy of the principals charged with leading and managing the 24 Mississippi middle schools that received low accountability scores suggest any connection to the ratings? Based on the findings of the investigation, it may be concluded that the self efficacy of the principals charged with leading and managing the middle schools that received low accountability scores didn’t appear to have any connection to the ratings. Also, neither the principals’ certification credential levels nor teaching endorsements appeared to be factors.

Credit scoring is critical for banks to evaluate new loan applications and monitor existing customers. Machine learning has been extensively researched for this case; however, the adoption of machine learning methods is minimal in financial risk management. The primary reason is that algorithms are viewed as ‘black box models’ and cannot satisfy regulatory requirements. While deep learning methods such as LSTM have been evaluated for behavioural credit scoring based on performance, research has not holistically evaluated these models on performance and explainability. To answer the research question, How can traditional machine learning and deep learning methods conform with regulatory guidelines for explainable artificial intelligence (XAI), and are they preferable to benchmark methods? this thesis used a public customer credit card dataset to compare the performance and explainability of machine learning and deep learning models against the benchmark statistical model linear regression. Model performance was evaluated using ROC-AUC, accuracy, Brier scores, F1 scores and the G-mean. The McNemar test evaluated whether, through pairwise comparison, the model performances were statistically different. The models were then evaluated on whether local and global explanations could be ascertained using feature/permutation importance and SHAP. The results found that neither the machine learning model, XGBoost, nor the deep learning model, LSTM, produced a statistically superior performance from the benchmark model. While there were performance improvements, only the machine learning model using post-hoc methods could produce local and global explanations. Given the strict regulatory environment, it is understandable that banks are hesitant to implement machine learning or deep learning models that lack the adequate levels of explainability regulators require.

Studien tar avstamp i praktikens frågor. Svensklärare och elever på gymnasieskolan har intervjuats för att skönja uppfattningar och erfarenheter om bedömningsmatrisers tillämpning i undervisningen. Studiens resultat har diskuterats ur ett OTL-perspektiv. Tre lärare deltog i studien och enskilda semistrukturerade intervjuer genomfördes med samtliga. Sex elever intervjuades i fokusgrupper uppdelade om tre elever i varje.  Resultatet presenteras genom tematisk analys av transkriberade intervjuer.  Studiens resultat visar att lärarna och eleverna delar liknande uppfattningar om och av bedömningsmatrisers tillämpning i undervisning. Bedömningsmatriser ansågs ha både för- och nackdelar som är av betydande roll dels för lärarens yrkesutövning, dels för elevens lärande. Bedömningsmatrisens största utmaning sågs av både lärare och elever som för abstrakt för att eleverna ska förstå skillnaden i kvaliteter mellan de olika nivåerna. Den största möjligheten med matrisanvändning uppfattades vara en tydlig kartläggning av elevens kunskaper. / The study is based on practice questions. Swedish teachers and students in upper secondary school have been interviewed to discern perceptions and experiences about the application of scoring rubrics in teaching. The results of the study have been discussed from an OTL perspective. Three teachers participated in the study and individual semi-structured interviews were conducted with all of them. Six students were interviewed in focus groups divided into three students in each group. The result of the study is presented through thematic analysis of transcribed interviews. The results of the study show that teachers and students share similar views about and of the application of scoring rubrics in teaching. Scoring rubrics were considered to have both advantages and disadvantages that are of significant role for both the teacher's professional practice and for the student's learning. The biggest shortcoming of the scoring rubric was seen by both teachers and students as too abstract for the students to understand the difference in qualities between the different levels. The biggest advantage of the rubric was perceived to be a clear mapping of the student's knowledge.

Kamratbedömning kan ha en positiv effekt på elevers lärande. För att uppnå denna positiva effekt måste elever kunna ge återkoppling till sina kamrater och även ta emot återkoppling från sina kamrater. Elever måste också kunna använda den återkoppling som de får. Återkoppling kan ges på fyra olika nivåer, uppgiftsnivå, processnivå, självregleringsnivå och personnivå. Att använda elever som lärande resurser för varandra, är en av nyckelstrategierna inom formativ bedömning. Fokus på formativ bedömning och användning av den samma har ökat inom skolan. Det, tillsammans med att det inte finns många studier som undersöker formativ bedömning med ett ämnesinnehåll i svensk kontext, är exempel på vad som ligger till grund för studien. Det övergripande syftet med denna studie var att öka kunskapen om hur kamratbedömning genomförs i naturvetenskap på mellanstadiet. Fokus riktades dels mot hur lärare implementerar kamratbedömning, men framförallt mot hur elever ger varandra återkoppling. För att undersöka hur och på vilka nivåer som 11-åriga elever ger varandra återkoppling, genomförde elever kamratbedömning i mindre grupp efter att enskilt ha besvarat uppgifter med ett naturvetenskapligt innehåll. Insamling av data skedde i flera steg. När läraren gav eleverna instruktioner kring bedömningsmatrisen, uppgifterna och den kamratbedömning som de skulle genomföra, observerades och videofilmades detta. Elevernas kamratbedömningssamtal i grupp, videofilmades också. Därefter intervjuades eleverna enskilt.Studiens resultat visade att eleverna fick återkoppling på olika nivåer av sina kamrater. Den återkoppling som var mest förekommande, var på uppgiftsnivå. Det fanns exempel på återkoppling på andra nivåer också i materialet, samt exempel där det inte förekom någon återkoppling alls. När eleverna bedömde varandras svar, fokuserade de på mängden naturvetenskapliga begrepp i kamraternas svar. Slutsatser som dragits utifrån studiens resultat var bland andra att elever behöver träna både på kamratbedömning och på att ge användbar återkoppling, samt att lärare bör vara medvetna om att deras instruktioner är betydelsefulla när det gäller hur kamratbedömningen faller ut. En annan slutsats är att kamratbedömning med ämnesinnehåll, kräver både bedömarfärdighet och ämneskunskap. / Peer assessment may have a positive effect on student learning. In order to have these positive effect students have to be able to give feedback to their peer and also to receive feedback from their peer. The students also have to be able to use the given feedback. Using students as learning resources to each other is a key strategy within formative assessment. The feedback that is given can be directed at four different levels, task level, process level, self-regulation level and self-level. Using peers as learning resources, is one of the key strategies within formative assessment. Focus on formative assessment and the use of it, have increased within school. That, and the fact that there are not many studies that examines formative assessment with science content, are a part of the background to the conducted study.The overall purpose of the study was to contributed to the field concerning peer assessment in science with 11-year old students. The focus was partly how teachers implement peer assessment, but above all how students give feedback. In order to examine how and in what different levels student give each other feedback, students conducted peer assessment in small groups, after they had answered questions concerning science. Data collection was done in the following steps. At first, the teacher gave the students instructions about the tasks, the scoring rubric, how to assess and so on. While doing that, the teacher was observed and video recorded. Next step concerned the students. They answered the questions and then had peer assessment in small groups. This was also video recorded. Later on the students were interviewed. The students received different kinds of feedback, mostly at task level, from their peer. Some examples of feedback at other levels were also discovered. When students assessed each other’s answers, they looked at the amount of science concepts that were used in their peer’s answers. The results suggest that students need to practice peer assessment, how to give useful feedback and that the teacher need to be aware of that given instruction may turn out differently from what is expected. The results also show that students believe that peer assessment is useful and that feedback is given. This was contradictory from what was seen in some of the observations of the peer assessment. Conclusions are for example that students need both assessment skills and scientific knowledge.

Syftet med denna magisteruppsats är att undersöka elevers uppfattningar och upplevelser av bedömningsmatriser i svenska. Uppsatsen är ett led i ett skolutvecklingsprojekt om kunskapssyn, bedömning och IUP, Individuell utvecklingsplan, på Fäladsgården, en 6-9-skola, i Lund . Med undersökningen vill jag få fördjupad förståelse för hur eleverna uppfattar och använder matriser och vad de anser sig behöva för att uppfatta lärarresponsen som god. Elva kvalitativa intervjuer med elever i skolår sju och nio genomfördes under december 2008 och januari 2009. Det är dessa elevers uppfattningar som tillsammans med litteratur i ämnet ligger till grund för arbetet.Resultatet av undersökningen visar att elevernas uppfattningar om bedömningsmatrisen i grunden är positiva, men att de efterfrågar kompletterande kommentarer till den. De önskar också mer tid för att bli väl förtrogna med innebörden i matrisen. Uppsatsens slutsatser, som har genererat en teori om aktivitetens betydelse och en arbetsmodell, kan väl peka ut inriktningen på det fortsatta utvecklingsarbetet på skolan även i andra ämnen än svenska och på andra skolor som använder matriser för ökat lärande. / The purpose of this essay is to examine pupils’ perceptions and experiences of scoring rubrics in Swedish. The essay is part of a of school development project focusing on knowledge, assessment and on the Individual Development Plan, at Fäladsgården, a compulsory school, in Lund, Sweden. In the survey, I want to achieve a deeper understanding of how pupils perceive and use scoring rubrics, and what they need to obtain a constructive feedback.Eleven qualitative interviews with pupils in school year seven and nine were carried out in December 2008 and January 2009. These pupils’ views are the basis for the paper.The results of the survey show that the pupils’ perceptions of the scoring rubrics are basically positive, but that they are asking for additional comments on it. They also want more time to familiar with the meaning of the scoring rubrics. The conclusions of this essay, which have generated a theory of the importance of the activity and a working model, may well indicate the continued development of this particular school and other schools, which use scoring rubrics for enhanced learning.

Malware is still one of the most prominent vectors through which computer networks and systems are compromised. A compromised computer system or network provides data and or processing resources to the world of cybercrime. With cybercrime projected to cost the world $6 trillion by 2021, malware is expected to continue being a growing challenge. Statistics around malware growth over the last decade support this theory as malware numbers enjoy almost an exponential increase over the period. Recent reports on the complexity of the malware show that the fight against malware as a means of building more resilient cyberspace is an evolving challenge. Compounding the problem is the lack of cyber security expertise to handle the expected rise in incidents. This thesis proposes advancing automation of the malware static analysis and detection to improve the decision-making confidence levels of a standard computer user in regards to a file’s malicious status. Therefore, this work introduces a framework that relies on two novel approaches to score the malicious intent of a file. The first approach attaches a probabilistic score to heuristic anomalies to calculate an overall file malicious score while the second approach uses fuzzy hashes and evidence combination theory for more efficient malware detection. The approaches’ resultant quantifiable scores measure the malicious intent of the file. The designed schemes were validated using a dataset of “clean” and “malicious” files. The results obtained show that the framework achieves true positive – false positive detection rate “trade-offs” for efficient malware detection.

Магистерская диссертация посвящена вопросам оценки кредитного риска заемщиков – физических лиц. Целью исследования является разработка методического подхода по оценке кредитоспособности заемщиков как направления совершенствования управления кредитным риском. В работе сделан вывод о том, что повышение качества кредитного портфеля банка способствует повышению конкурентоспособности банка в условиях высокой конкуренции и нормативных требований Банка России. / Master thesis is devoted to the assessment of credit risk of individual borrowers. The aim of the study is the development a methodical approach to assess the creditworthiness of borrowers as a way to improve credit risk management. The work concluded that improving the quality of the credit portfolio of a bank contributes to improving competitiveness of bank in conditions of high competition and regulatory requirements of the Bank of Russia.