Saara Hyvönen’s research while affiliated with Helsinki School of Economics and other places

We investigate two recommendation approaches suitable for online multimedia sharing services. Our first approach, UserRank, recommends items by global interestingness irrespective of user preferences and is based on the analysis of ownership and evaluation link structure. We also present a personalized interestingness algorithm that combines UserRank with collaborative filtering which enables a single parameter to control the degree of personalization in the recommendations. Our initial results from an informal user study are encouraging.

The Finnish academic marketing research and the marketing research made in practice can be clearly distinguished. However, there are many efforts to narrow the gap and marketing information systems are emerging which constitute a positive factor in the use of analytical marketing models and methodology in business decisions. The use of marketing concept in extending to the areas of of industrial marketing, marketing of services and nonprofit and government marketing. There is also a growing interest in strategic marketing planning. This paper presents to a large extent recent efforts in the field of academic marketing research in Finland. The natural starting point is the distribution research, which has strong traditions in all Nordic countries. In Finland the competitive, open market has been gradually replaced by four vertical marketing systems or blocks, which account for about 90 per cent of the grocery trade at the retail level. There is also a tendency towards integration in the specialty goods and industrial goods trade. The result has been increased competition between systems. There is the problem how to manage a whole interorganizational system, which consists of numerous enterprises of varying sizes. These problems have been recognized by distribution channel managers, government agencies and academicians. It is suggested that the behavioral and economic aspects are both important in the research of interorganizational coordination. The performance of the entire vertical marketing system is dependent on each of its subunits. The productive capacities of subunits constitute the most challenging areas in the strategic planning of marketing systems. Among the decision areas of marketing mix pricing is perhaps the least analysed with behavioral concepts and methodology. Cost-based pricing strategies are still dominant in Finnish business life. Different ways to measure consumer price perception and price sensitivity are discussed in a brand choice situation. The problems in applying simulation models in the decision making of marketing management is next considered by a representative of managers. Finally, the possibilities of descriptive modeling 1n organ1zat1onal buying analysis are illustrated and the results of a Finnish intensive study on the buying process of industrial components are reported.

Category management (CM) has been promoted as a joint business process between manufacturers and retailers to produce enhanced business results by focusing on delivering better value to end consumers. But do the business partners involved in CM really recognize which particular CM arrangements generate these results, what possibilities do they have to influence CM collaboration, and does CM most benefit the more powerful channel members? To address these issues, our aim is to explore the effects of manufacturers' influence strategies (consumer-oriented activities, information exchange supporting category plan implementation, and incentives) and control over CM on category performance. The data consist of 206 responses to a survey of food manufacturers in Finland and Sweden. Our findings indicate that plan implementation and control enhance most manufacturers' category performance, and that control is contingent on a manufacturer's category position. Category leaders, and to a smaller extent, category supporters tend to have most influence in CM collaboration. Finnish and Swedish manufacturers also had significant differences in their influence strategy–category performance profiles.

This book explores dialects and social differences in language computationally, examining topics such as how (and how much) linguistic differences impede intelligibility, how national borders accelerate and direct change, how opinion and hearsay shape perceptions of language differences, the role of intonation (melody), the differences between variation in pronunciation and vocabulary, and techniques for recognising structure in larger collections of linguistic data. The computational investigations engage more traditional work deeply, and a panel discussion focuses on the opportunities and risks of pursuing humanities research using computational science. There is also an introduction that attempts to sketch perspectives from which to approach the individual contributions.

The purpose of this paper is to analyze suppliers' roles in category management (CM) in the context of Finnish and Swedish supplier–retailer relationships. Using data from a survey of a sample of Finnish and Swedish suppliers, the study shows that the concept of CM is well known among both Finnish and Swedish suppliers, and that most of them have experience of CM as a key operational business process. The study also shows that, in general, larger suppliers in both Finland and Sweden have a relatively strong role in CM collaboration, whereas smaller suppliers (especially in Finland) are more likely to have a lesser role in CM collaboration. Suppliers with a strong role in CM collaboration are generally more capable of influencing individual CM tactics than suppliers with an equal role to competitors or those with no role (both in Finland and in Sweden). The study also finds that the effects of CM are perceived more positively among suppliers who have influence in decision-making regarding CM collaboration than among those who have no role in such decision-making. However, the opinions of suppliers who have no role in CM collaboration are neutral, rather than being especially negative. Qualitative studies, including face-to-face discussions with managers representing manufacturers who are more or less excluded from CM collaboration, might provide a more thorough understanding of CM from their perspective.

Component models such as factor analysis can be used to analyse spatial distributions of a large number of different features – for instance the isogloss data in a dialect atlas, or the distributions of ethnological or archaeological phenomena – with the goal of finding dialects or similar cultural aggregates. However, there are several such methods, and it is not obvious how their differences affect their usability for computational dialectology. We attempt to tackle this question by comparing five such methods using two different dialectological data sets. There are some fundamental differences between these methods, and some of these have implications that affect the dialectological interpretation of the results. © Edinburgh University Press and the Association for History and Computing 2009.

A matrix decomposition expresses a matrix as a product of at least two factor matrices. Equivalently, it expresses each column of the input matrix as a linear combination of the columns in the first factor matrix. The interpretability of the decompositions is a key issue in many data-analysis tasks. We propose two new matrix-decomposition problems: the nonnegative CX and nonnegative CUR problems, that give naturally interpretable factors. They extend the recently-proposed column and column-row based decompositions, and are aimed to be used with nonnegative matrices. Our decompositions represent the input matrix as a nonnegative linear combination of a subset of its columns (or columns and rows). We present two algorithms to solve these problems and provide an extensive experimental evaluation where we assess the quality of our algorithms' results as well as the intuitiveness of nonnegative CX and CUR decompositions. We show that our algorithms return intuitive answers with smaller reconstruction errors than the previously-proposed methods for column and column-row decompositions.

We investigate two recommendation approaches suitable for online multimedia sharing services. Our first approach, UserRank, recommends items by global interestingness irrespective of user preferences and is based on the analysis of ownership and evaluation link structure. We also present a personalized interestingness algorithm that combines UserRank with collaborative filtering which enables a single parameter to control the degree of personalization in the recommendations. Our initial results from an informal user study are encouraging.

Binary data appears in many spatial applications such as dialectology and ecology. We demonstrate that a simple Bayesian modeling approach can be used in pre-processing large spatial data sets with missing or uncertain data. Our experiments on real and synthetic data show that conducting the pre-processing phase before applying conventional data mining methods, such as PCA, clustering or NMF, improves the results significantly.