På uppdrag av sportmediakoncernen ESMG undersöker detta arbete två frågeställningar: Hur kan man utveckla och driftsätta ett rekommendationssystem för nyhetsartiklar? Vilka föroch nackdelar finns med ett eget system jämfört med tredjepartssystem? Arbetet använder Polyas fyra steg som undersökningsmetod, där de fyra stegen anpassas och appliceras på detta projekt. För att kunna besvara den första frågeställningen, skapas initialt en kravspecifikation, som ligger till grund för rekommendationssystemets funktionella och icke-funktionella krav. Utifrån kravspecifikationen, görs en initial fallstudie av Amazon Web Services (AWS), där lämpliga verktyg och tjänster väljs, följt av utformning av en arkitektur för rekommendationssystemet. I en fallstudie av Hockeysveriges webbplats, implementeras sedan arkitekturen med hjälp av AWS och några andra verktyg, som Google Tag Manager och Numeri. Slutligen utvärderas arbetet för kravuppfyllnad. För att kunna besvara den andra frågeställningen, görs en summativ utvärdering av ett antal olika tredjepartssystem för rekommendationer. Genom analys av tredjepartssystemens respektive webbplatser, tas listor på föroch nackdelar fram, ackompanjerat med korta beskrivningar av tjänsterna. Resultaten av den första frågeställningen är en lösning, som visar hur man i praktiken kan utveckla och driftsätta ett rekommendationssystem för nyhetsartiklar. Genom en detaljerad beskrivning alla delar av utvecklingen, fungerar resultaten som en konkret guide i skapande av rekommendationssystem med moderna verktyg. Med avseende på arbetets andra frågeställning, visar resultaten att den stora skillnaden mellan ett egenbyggt system och tredjepartssystem ligger i flexibiliteten, men att ett eget system kommer med mer ansvar, fler beroenden och utan annan funktionalitet som statistik, vilket ofta ingår i tredjepartssystem. / On behalf of the corporate group ESMG, this thesis examines two research questions: How can one develop and deploy a custom recommender system for news articles? What are the pros and cons of having a custom system, compared to third-party systems? The thesis utilizes Polya's four steps as its research method, where the four steps are adapted and applied to this particular project. In order to answer the first research question, an initial requirements specification is created, which provides the basis for the recommender system's functional and non-functional requirements. Based on the requirement specification, an initial case study of Amazon Web Services (AWS) is performed, where appropriate tools and services are selected, followed by the design of an architecture for the recommender system. In a case study of ESMG:s website Hockeysverige, the architecture is then implemented, using AWS and some other necessary tools, such as Google Tag Manager and Numeri. Finally, the implementation is evaluated with respect to requirement compliance. To answer the second research question, a summative evaluation of a number of different third-party recommender systems is performed. By analyzing the third-party systems' websites, a list of pros and cons is presented, accompanied by a brief description of the service. The results of the first research question, is a solution which illustrates how one can, in practice, implement a news recommender system. Through a detailed description of all aspects of development, the results function as a guide in creating recommendation systems using modern tools. With regard to the second research question, the results show that the major difference between a custom system and third-party systems, lies in the flexibility, but a custom system brings more responsibility, more dependencies, and no other functionality, such as statistics, which is often part of third-party systems.

In content-based image retrieval (CBIR) systems, there are two general types of search: target search and category search. Unlike queries in traditional database systems, users in most cases cannot specify an ideal query to retrieve the desired results for either target search or category search in multimedia database systems, and have to rely on iterative feedback to refine their query. Efficient evaluation of such iterative queries can be a challenge, especially when the multimedia database contains a large number of entries, and the search needs many iterations, and when the underlying distance measure is computationally expensive. The overall processing costs, including CPU and disk I/O, are further emphasized if there are numerous concurrent accesses. To address these limitations involved in relevance feedback processing, we propose a generic framework, including a query model, index structures, and query optimization techniques. Specifically, this thesis has five main contributions as follows. The first contribution is an efficient target search technique. We propose four target search methods: naive random scan (NRS), local neighboring movement (LNM), neighboring divide-and-conquer (NDC), and global divide-and-conquer (GDC) methods. All these methods are built around a common strategy: they do not retrieve checked images (i.e., shrink the search space). Furthermore, NDC and GDC exploit Voronoi diagrams to aggressively prune the search space and move towards target images. We theoretically and experimentally prove that the convergence speeds of GDC and NDC are much faster than those of NRS and recent methods. The second contribution is a method to reduce the number of expensive distance computation when answering k-NN queries with non-metric distance measures. We propose an efficient distance mapping function that transfers non-metric measures into metric, and still preserves the original distance orderings. Then existing metric index structures (e.g., M-tree) can be used to reduce the computational cost by exploiting the triangular inequality property. The third contribution is an incremental query processing technique for Support Vector Machines (SVMs). SVMs have been widely used in multimedia retrieval to learn a concept in order to find the best matches. SVMs, however, suffer from the scalability problem associated with larger database sizes. To address this limitation, we propose an efficient query evaluation technique by employing incremental update. The proposed technique also takes advantage of a tuned index structure to efficiently prune irrelevant data. As a result, only a small portion of the data set needs to be accessed for query processing. This index structure also provides an inexpensive means to process the set of candidates to evaluate the final query result. This technique can work with different kernel functions and kernel parameters. The fourth contribution is a method to avoid local optimum traps. Existing CBIR systems, designed around query refinement based on relevance feedback, suffer from local optimum traps that may severely impair the overall retrieval performance. We therefore propose a simulated annealing-based approach to address this important issue. When a stuck-at-a-local-optimum occurs, we employ a neighborhood search technique (i.e., simulated annealing) to continue the search for additional matching images, thus escaping from the local optimum. We also propose an index structure to speed up such neighborhood search. Finally, the fifth contribution is a generic framework to support concurrent accesses. We develop new storage and query processing techniques to exploit sequential access and leverage inter-query concurrency to share computation. Our experimental results, based on the Corel dataset, indicate that the proposed optimization can significantly reduce average response time while achieving better precision and recall, and is scalable to support a large user community. This latter performance characteristic is largely neglected in existing systems making them less suitable for large-scale deployment. With the growing interest in Internet-scale image search applications, our framework offers an effective solution to the scalability problem.

Bakgrund och problemformulering: Hållbarhetsrapporter blir allt viktigare för företag över hela världen till följd av samhällets fruktan och oro för miljöförstöringar, växthuseffekten, klimatförändringar och även mänskliga rättigheter där metoder rörande ett socialt ansvar uppmuntrats. Skogen framstår som viktig för hållbara material och produkter då den redogörs som väsentlig i klimatomställningen. Ett problem som har påvisats utifrån tidigare studier är kvaliteten på rapporterna. Kritik gentemot rapporterna har uppkommit där det föreligger en avsaknad av både trovärdighet och relevans.  Syfte: Syftet med denna studie är att undersöka kvaliteten i hållbarhetsrapporter i form av trovärdighet och relevans och få en djupare kunskap om kvaliteten i företagens hållbarhetsarbete. Vidare jämförs förändringar i kvaliteten i de valda företagens hållbarhetsrapporter mellan åren 2017-2021.  Forskningsfråga: Har kvaliteten i hållbarhetsrapporter utifrån trovärdighet och relevans inom skogsbranschen förändrats mellan åren 2017-2021? Metod: Denna studie är en komparativ och en longitudinell studie som bygger på en kvalitativ forskningsdesign med kvantitativa inslag i form av en innehållsanalys. För att undersöka kvaliteten i företagens hållbarhetsrapporter har ett kodningsschema med ett poängsystem tillämpats som baseras på trovärdighet och relevans.   Analys och slutsats: Resultatet visar en ökning av kvaliteten i hållbarhetsrapporterna mellan åren 2017-2021, där både trovärdigheten och relevansen har ökat. Hållbarhetsrapporternas trovärdighet översteg relevansen i de undersökta rapporterna. / Background and problem formulation: Sustainability reports are becoming increasingly important for companies all over the world as a result of society's fear and concern about environmental degradation, the greenhouse effect, climate change and also human rights where social responsibility practices have been encouraged. The forest appears to be important for sustainable materials and products as it is described as essential in climate change. A problem that has been demonstrated based on previous studies is the quality of the reports. Criticism of the reports has arisen where there is a lack of both credibility and relevance. Purpose: The purpose of this study is to investigate the quality of sustainability reports in terms of credibility and relevance and to gain a deeper knowledge of the quality of the companies' sustainability work. Furthermore, changes in the quality of the selected companies' sustainability reports between the years 2017-2021 are compared. Research question: Has the quality of sustainability reports based on credibility and relevance in the forest industry changed between the years 2017-2021? Method: This study is a comparative and a longitudinal study based on a qualitative research design with quantitative elements in the form of a content analysis. To examine the quality of the companies' sustainability reports, a coding scheme with a scoring system based on credibility and relevance has been applied. Results and conclusion: The result shows an increase in the quality of the sustainability reports between the years 2017-2021, where both credibility and relevance have increased. The credibility of the sustainability reports exceeded the relevance of the investigated reports.

Making inferences in communication is a highly context-dependent endeavor. Previous research found cultural variations for context-sensitivity as well as for communication comprehension. However, the relative impact of culture and context-sensitivity on communication comprehension has not been investigated so far. The current study aimed at investigating this interplay and tested 4- and 6-year-old children from Germany (n = 132) and China (n = 129). Context-sensitivity was measured with an adapted version of the Ebbinghaus illusion. In this task, children have to discriminate the size of two target circles that only appear to be of similar size due to context circles surrounding the target circles. As expected, performance scores indicated higher degrees of context-sensitivity in Chinese compared to German children and that 6-year-olds were more context-sensitive than 4-year-olds. Further, in an object-choice communication-comprehension task, children watched videos with puppets performing everyday activities (e.g., pet care) and had to choose between two options (e.g., dog or rabbit). A puppet expressed what she wanted either directly (“I want the rabbit”) or indirectly (“I have a carrot”). The children had to choose one option to give to the puppet. In both cultures, 6-year-olds outperformed 4-year-olds and children understood direct communication better than indirect communication. Culture was found to affect children’s processing speed of direct communication. Moreover, culture influenced children’s context-sensitivity while contextsensitivity influenced children’s accuracy in the indirect (but not the direct) communication task. These findings demonstrate that taking context into account is especially important when we are confronted with indirect communication.

In a society where digitalization thrives, opportunities for both conducting e-commerce and promoting it are enhanced. A common marketing strategy that has shown to be efficient is collaborating with Social Media Influencers. It is widely known that social media influencers can greatly help e-commerce companies get noticed. However, it has remained a problem figuring out what an appropriate influencer is for one’s business purposes and marketing objectives.  This study has gathered empirical data by conducting qualitative studies, including interviews with four e-commerce companies. The researchers followed a deductive approach in their investigation. The interviews have helped identify patterns related to the research questions and have generated new insights in the field. The research provides a deeper understanding of which parameters are suitable for specific purposes. Additionally, the thesis highlights the interconnections between these parameters to elaborate an overall understanding of the strategy. The potential risks associated with the strategy are also addressed. This thesis highlights why certain social media influencers are considered more suitable than others in the perspective of an e-commerce business.

No description available.

No description available.

No description available.

No description available.

Yes / This paper uses panel cointegration with a corresponding vector error correction model (VECM) to investigate the changes in the value relevance of accounting information before and after the mandatory adoption of IFRS in Germany and the UK under three different valuation models. First, a basic Ohlson model, where our results indicate that despite the value relevance of the book values of equity has declined, it has been replaced by the increasing prominence of earnings in both Germany and the UK after the switch to the IFRS. Second, a modified model, which shows that the incremental value relevance of both earnings and book values are considerably higher in the long term for firms in the UK than in Germany. Third, a simultaneous addition of accounting and macroeconomic variables in an extended model, which indicates a significant rise in the relative predictive power of the book value of equity in the UK compared with the more noticeable impact on the value relevance of earnings in Germany. Collectively, the results of these models indicate that: (i) the explanatory power of linear equity valuation models is higher in UK than in the Germany, (ii) a long-run Granger-causal relationship exists between accounting variables and share prices in common law countries like the UK. Nevertheless, the implications of our findings lie in the knowledge that the potential costs of switching to the IFRS is completely nullified within three years by the benefits arising from a reduction in information asymmetry and earning mismanagement among firms which are listed on the stock exchanges of both common law and code law-based EU countries.