Hide menu

Examensarbeten och uppsatser / Final Theses

Framläggningar på IDA / Presentations at IDA


Se även framläggningar annonserade hos ISY och ITN i Norrköping / See also presentations announced at ISY and ITN in Norrköping (in Swedish)

If nothing is stated about the presentation language then the presentation is in Swedish.


WExUpp - kommande framläggningar
2018-06-18 - SaS
User-Based Predictive Caching of Streaming Media
Fredrik Håkansson
Avancerad (30hp)
kl 10:15, John von Neuman (In English)
[Abstract]
Streaming media is a growing market all over the world and sets a strict requirement on mobile connectivity. The foundation for a good user experience when supplying a stream- ing media service on a mobile device is to ensure that the user can access the requested content. Due to the varying availability of mobile connectivity measures has to be taken to remove as much dependency as possible on the quality of the connection. This thesis investigates the use of a Long Short-Term Memory machine learning model for predicting a future geographical location for a mobile device. The predicted location in combination with information about cellular connectivity in the geographical area is used to schedule prefetching of music content in order to improve user experience and to reduce mobile data usage. The Long Short-Term Memory model suggested in this thesis achieves an accuracy of 85.15% averaged over 20000 routes and the predictive caching managed to retain user experience while decreasing the amount of data consumed.
2018-06-18 - HCS
MDM Utvärdering
Fanny Lindmark
Avancerad (30hp)
kl 10:30, Alan Turing (På svenska)
[Abstract]
Connectivity provided by technology and liberation of trade have led to a globalization of organizations, affecting supply chains to expand in complexity. As a result, many organizations today have challenges of managing information in a consistent manner throughout a complex system environment. This study aims to identify the most valuable attributes of a solution for managing master data, in an efficient and consistent manner, across an extended and complex supply chain. Master data, such as products, customers and suppliers, can be defined as valuable core business information, since it is vital for supporting business operations. A requirements elicitation was performed, including interviews conducted internally with employees at IFS and externally with customers. Furthermore, a requirements analysis resulted in a specification of requirements including the most desirable attributes of a future Master Data Management (MDM) solution. Five main themes of the attributes were identified; architecture, availability and integration, governance, user interface and lifecycle management. The study contributes to the area of research, by identifying challenges and valuable attributes to consider when developing or investing in a solution for MDM.
2018-06-18 - STIMA
Optimized trade execution with reinforcement learning
Olle Dahlén, Axel Rantil
Avancerad (30hp)
kl 13:15, John von Neumann (På svenska)
[Abstract]
In this thesis, we study the problem of buying or selling a given volume of a financial asset within a given time horizon to the best possible price, a problem formally known as optimized trade execution. Our approach is an empirical one. We use historical data to simulate the process of placing artificial orders in a market. This simulation enables us to model the problem as a Markov decision process (MDP). Given this MDP, we train and evaluate a set of reinforcement learning (RL) algorithms all with the objective to minimize the transaction cost on unseen test data. We train and evaluate these for various instruments and problem settings, such as different trading horizons.
Our first model was developed with the goal to validate results achieved by Nevmyvaka, Feng and Kearns (2006) and it is thus called NFK. We extended this model into what we call Dual NFK, in an attempt to regularize the model against external price movement. Furthermore, we implemented and evaluated a classical RL algorithm, namely Sarsa(λ) with a modified reward function. Lastly, we evaluated proximal policy optimization (PPO), an actor-critic RL algorithm incorporating neural networks in order to find the optimal policy. Along with these models, we implemented five simple baseline strategies with various characteristics. These baseline strategies have partly been found in the literature and partly been developed by us, and are used to the evaluate the performance of our models.
We achieve results on par with those found by Nevmyvaka, Feng and Kearns (2006), but only for a few cases. Furthermore, dual NFK performed very similar to NFK, indicating that one can train one model (for both the buy and sell case) instead of two for the optimized trade execution problem. We also found that Sarsa(λ) with a modified reward function performed better than both these models, but is still outperformed by baseline strategies for many problems. Finally, we evaluated PPO for one problem setting and found that it outperformed even the best of the baseline strategies and models, showing promise for deep reinforcement learning methods for the problem of optimized trade execution.
2018-06-18 - SaS
Usability of a Business Software Solution for Financial Followup Information of Service Contracts
Therese Borg
Avancerad (30hp)
kl 13:15, Alan Turing (På svenska)
[Abstract]
Enterprise Resource Planning systems have been available since the 1990s and come with several business benefits for the users. One of the major advantages is improved decision-making through current and accessible information about strategical, tactical and operational levels of the organization. Although several Enterprise Resource Planning system vendors provide several features for contract management, more decision support regarding the total profitability of service contracts is desired by the customers. Estimating the total profitability of service contracts is a challenging task for all service providers and implies a lot of manual data processing by the contract manager.
This master’s thesis is conducted in collaboration with IFS World Operations AB and aims to investigate how functionality for budget and forecasting of the profitability of service contracts can be designed to be usable in terms of effectiveness. The implementation was performed iteratively and the resulting prototypes were evaluated and refined throughout the project. The final high-tech prototype for budgeting of service contracts was evaluated using the task success rate in conjunction with the System Usability Scale to assess how well the system conformed
to the needs of the users. The study revealed that two of the key characteristics of financial follow-up information of service contracts is the support of creating a budget and graphical visualizations of both budgeted and actual values. The final usability evaluation indicated that the developed functionality was usable in terms of effectiveness and has an overall usability clearly above the average.
2018-06-18 - HCS
Static Code Analysis of Progressive Web Applications in Angular
Fabian Johannsen
Avancerad (30hp)
kl 13:15, Herbert Simon (På svenska)
[Abstract]
Web applications have a common code base across multiple platforms, but have previously lacked some core features compared to native applications. However, recent web technology advancements have, in terms of functionality and user experience, reduced the gap between the two development approaches. Applications that leverages these technology advancements are dubbed progressive web applications, or PWA. This thesis explores the concepts of PWA and how it, in terms of code complexity, affects an Angular web application. The results show that implementing considered PWA features does not excessively increase the size of an application and that the overall added complexity is low. The complexity of PWA lies in all the new technology concepts, which are probably unfamiliar to most developers. To reduce this complexity, automated PWA tooling shows great promise, and using Angular PWA tooling when building Angular applications seems to minimize this complexity.
2018-06-18 - SaS
Användning av prototyper som verktyg för kravhantering i agil mjukvaruutveckling
Linn Abrahamsson, Peter Melin Wenström
Avancerad (30hp)
kl 15:00, (På svenska)
[Abstract]
Kravhantering vid agil mjukvaruutveckling är en utmaning som alltfler företag ställs inför och för detta finns många verktyg att välja mellan. Prototyper, modeller som liknar tilltänkta slutprodukter, är ett sådant verktyg som används för att inhämta viktig information om det som ska utvecklas. För att beskriva hur lik en prototyp är den tilltänkta slutprodukten används begreppet verklighetsfaktor. Studiens syfte är dels att öka kunskapen kring prototypanvändning i agil mjukvaruutveckling, dels att undersöka vilken effekt en prototyps verklighetsfaktor har då prototyper används i diskussioner inom kravhantering. En fallstudie görs på företaget Exsitec där personal intervjuas angående prototypanvändning i mjukvaruprojekt. Två prototyper utvecklas sedan med låg respektive hög verklighetsfaktor och används som diskussionsunderlag i intervjuer. Studien visar att användning av prototyper i mjukvaruprojekt kan bidra till ökat förtroende hos kunder, förbättrad kommunikation med kunder och kan förenkla att uppnå konsensus mellan olika intressenter. Vidare kan de, beroende av hur de används, bidra till helhetsbilden av produkten och fungera som dokumentation. Studien påvisar även några, om än subtila, skillnader i den information som samlas in med hjälp av prototyper med låg respektive hög verklighetsfaktor. Hög verklighetsfaktor tycks medföra att fler krav samlas in, men göra respondenter mindre benägna att vilja komma med förslag på mer omfattande förändringar
2018-06-18 - SaS
Can the Google PageRank algorithm be used to analyze the communication within a company?
Henning Nåbo
Avancerad (30hp)
kl 15:00, Herbert Simon (In English)
[Abstract]
This paper describes the process of modeling and analysing corporate communication.
The data is taken from the Briteback enterprise communication application. Communication
through phone and chat is studied by creating communication graphs. Phone
communication is measured by number of calls and duration of calls, and chat communication
is measured by the minimum number of messages sent from one person to another.
These measurements are then used to calculate a communication score, different methods
are tested and a version using principal component analysis is chosen. Different centrality
measurements are performed on the graph model; eigenvector centrality, page rank and
betweenness centrality. These are all found to be useful in different ways; eigenvector
centrality fits the data best, page rank is easy to understand and can be adapted for
different situations, and betweenness points out users in critical positions in the communication
graphs. Personalized Page Rank focused on users or a group of users is tested
and shows potential to be of use for social network service companies in many different
ways such as when ordering search results or when suggesting new members to a chat
channel.
2018-06-19 - SaS
Förbättring av AR-visualiseringen med Kalmanfiltrering och horisontbaserad orientering - för att förhindra båtar att gå på grund
Pontus Hero Ek
Avancerad (30hp)
kl 10:00, Alan Turing (På svenska)
[Abstract]
This thesis researched the possibility of improving the compass of smartphones as the earth’s magnetic field is not strong and is easily disturbed, either by the environment or technology. The compass is used in Augmented Reality (AR) when the AR visualization should correspond to a position on earth. The issue lies in oscillating input values to the compass that reduces the AR experience. To improve the AR experience without the use of external equipment, this work tried to both filter the incoming values with a Kalman filter and to know the direction by capturing an image with a horizon that was image processed. The Kalman filter achieved a reduction in incoming disturbances and the horizon was matched against a panorama image that was generated from 3D data. The thesis starts off with requirements and contents of AR and goes through the different approaches that begins with a LAS point cloud and ends in matching horizons with normalized cross-correlation. This thesis furthermore measures performance and battery drainage of the built application on three different smartphones that are nearly a year apart each. Drift was also measured as it is a common issue if there is no earthly orientation to correct itself unto, for instance the magnetometer. This showed that these methods can be used on OnePlus 2, Samsung Galaxy S7, and Samsung Galaxy S8, there is a steady performance and efficiency increase in each generation and that ARCore causes less drift. Furthermore this thesis shows the difference between a compass and a local orientation with an offset. The application that was made focused to work at sea but it was also tested on buildings with good results. The application also underwent usability tests that showed that the applied functionalities improved the AR-experience. The conclusion shows that it is possible to improve the orientation of smartphones. Albeit it can go wrong sometimes which is why this thesis also presents two ways to indicate that the heading is off.
2018-06-19 - AIICS
Evaluating Deep Learning Algorithms for Steering an Autonomous Vehicle"
Filip Magnusson
Avancerad (30hp)
kl 13:15, John von Neumann (In English)
[Abstract]
With self-driving cars on the horizon, vehicle autonomy and its problems is a hot topic. In this study we are using convolutional neural networks to make a robot car avoid obstacles. The robot car has a monocular camera, and our approach is to use the images taken by the camera as input, and then output a steering command. Usng this method the car is to avoid any object in front of it.

In order to lower the amount of training data we use models that are pretrained on ImageNet, a large image database containing millions of images. The model are then trained on our own dataset, which contains of images taken directly by the robot car while driving around. The images are then labeled with the steering command used while taking the image. While training we experiment with using different amounts of frozen layers. A frozen layer is a layer that has been pretrained on ImageNet, but are not trained on our dataset.

We find that a lower amount of frozen layers produces better results, and our best model achieved 81.19% accuracy on our test set. During a qualitative test the car avoid collisions 78.57% of the time.
2018-06-19 - HCS
Handledning för pusselspel med lärandet i fokus
Nikolas Spyrou
Grundnivå (16hp)
kl 14:15, Herbert Simon (På svenska)
[Abstract]
Det blir allt vanligare att spel används för att lära ut olika ämnen som matematik eller historia. De flesta och särskilt barn idag har tillgång till någon typ av surfplatta eller smartphone. Det är därför viktigt att dessa spel introduceras på rätt sätt för att fånga barnens uppmärksamhet. Hur ska en handledning för ett mobilt pusselmattespel som har lärandet i centrum utformas för att ge en positiv spelupplevelse? Detta är frågan som denna rapport har undersökt. Tangle Math är ett pusselspel som försöker få användaren att träna på huvudräkning. Det är viktigt att alla delar i spelet fungerar bra så att användaren får bästa möjliga spelupplevelse. För att uppnå detta måste vi förklara spelreglerna på ett sådant sätt så att alla förstår vad man ska göra utan att man får ör mycket hjälp. Denna rapport undersöker spelet i två iterationer. I iteration 1 har tester genomförts huvudsakligen med barn ifrån årskurs tre. För iteration 2 två har vuxna i åldern 19 till 30 använts. Testpersonerna spelade spelet och fyllde sedan i de tre olika enkäterna UEQ, IEQ och PANAS. Bearbetning av resultaten från första iterationen medförde ett flertal ändringar som ledde fram till spelets och handledningens slutversion. Ändringar har utförts kontinuerligt från testtillfälle till testtillfälle för att få användaren att lättare förstå målet samt spelets regler och samband. Förståelsen av korrelationen mellan färgerna och summan har stärkts genom att införa ordet färgsumma som antar färgen av det valda talet. När ordet summa ändrades till färgsumma behövde också målet bli tydligare så en text med beskrivning av målet lades till. Två cirklar och ett likhetstecken lades till i handledningen vilket resulterade i att målet blev tydligare. För att öka motivationen hos användaren har en räknare som berättar hur många fler rätta svar som krävs för att få se bilden eller för att få en medalj lagts till. Förutom medaljer får användaren beröm när olika delmål uppnåtts. En bild lades också in med krav för att uppnå nästa mål. För att spelaren på ett tydligt ska sätt veta vilka nivåer som är avklarade ändrades handledningen så att de avklarade nivåerna är ljusare medan de som ännu inte klarats av är mörkare. Olika meddelanden som hjälper användaren att bättre förstå spelets regler har lagts till. Ju fler kontinuerliga felförsök användaren gör desto mer hjälp får man. I handledningen har utropstecken (!) adderats. Användaren kan efter behov klicka på! för att få ytterligare hjälp. Sist, för att säkerställa att ingen avslutar spelet av misstag har ett meddelande som frågar om man har valt att avsluta spelomgången lagts till. Slutversionen validerades i iteration 2 och visade att användarnas spelupplevelse var positiv. Sammanfattningsvis, Tangle Math, ett spel med lärande i fokus, har ändrats mycket under arbetets gång, det har blivit lättare att förstå och mer inspirerande vilket användarnas uppskattning har visat. För spel som har lärandet i fokus är det viktigt att utvärdering utförs med olika målgrupper. På detta sätt utvidgas spelets funktionalitet och användningsområde. Det betyder till exempel att barn kan spela och samtidigt träna på ämnet och ger läraren en möjlighet att använda spelen som ett komplement till undervisningsplanen.
2018-06-21 - ADIT
Format Conversions and Query Rewriting for RDF* and SPARQL*
Jesper Eriksson, Amir Hakim
Grundnivå (16hp)
kl 09:00, Muhammad al-Khwarizmi (In English)
[Abstract]
Resource Description Framework (RDF) is a modern graph database model which has a proposed extension called RDF*. However, currently there are no tools available that allow for transitions between RDF and RDF*. In this thesis tools are developed that allow for conversions between these two formats. Tools for converting between RDF* and another graph database model called Property Graph has also been developed. Measurements of the memory usage and conversion times for the developed conversion tools were made. Tests were also performed to control that there was no data loss in the conversions. The results showed varying memory usage and conversion times for the different conversions and that some conversions were very difficult to check for data corruption. Complex testing tools would be required to make sure all conversions are made correctly.
2018-06-21 - HCS
Program för digitala prov
Marcus Jackson
Grundnivå (16hp)
kl 13:15, Herbert Simon (På svenska)
[Abstract]
Undersökning av hur ett generellt nedlåsningsprogram för digitala prov klarar av att möta de krav som ställs när det ska implementeras i en BYOD-miljö i en gymnasieskola.
2018-06-21 - ADIT
Result size calculation for Facebook's GraphQL query language
Tim Andersson
Grundnivå (16hp)
kl 14:00, Muhammad al-Khwarizmi (In English)
[Abstract]
GraphQL is a query language for web APIs and a service for carrying out server requests
to interact with the data from the API. Research show that even for simple GraphQL queries
both the size of the response object and the execution times to retrieve these objects may be
prohibitively large and that current implementations of the language suffers from this issue.
This thesis explores the implementation of an algorithm for calculating the exact size of the
response object from a GraphQL query, and the performance based evaluation of the implementation.
A proof of concept of a server using the implementation of the algorithm and subsequent tests of the
calculation times for particularly problematic queries sent to the server, show that the implementation
scales well and could serve as a way to stop these queries from executing.


Page responsible: Ola Leifler
Last updated: 2017-04-27