Browsing by Subject "self-organizing map"

Sort by: Order: Results:

Now showing items 1-4 of 4
  • Kordelin, Toni Juha (Helsingfors universitet, 2016)
    The importance of the computer and information technology has increased significantly with the introduction of the Internet. The technologies should not be reduced to selling and purchasing functions but should also be used as research and business intelligence tool, as well as information source. This study compares the performances of the Finnish and German sawmilling & paper industry by evaluating their public communication. The forest sector has been traditionally important for the economy of both countries and they belong to the global players. In conjunction with the performance analysis, a neural network tool is compiled, optimized and tested. It combines a study related categorization application and the Kohonen’s self-organizing map (SOM). The objective of the automated categorization is to lower manual work and to evaluate text contents more accurately – by taking also the possible future use into account. The research includes two interlinked approaches: web page and performance analysis. The web page analysis compares the design and content of the forest industry companies’ Internet sites. In addition, it is used to select and define the relevant enterprises for the performance analysis. 14 Finnish and 10 German sawmills, as well as 5 Finnish and 12 German paper and paperboard producers fulfilled the research requirements (data collection 2013). The performance analysis is implemented by a content analysis and SOM Tool. The word count data of the content analysis is sorted with a customised classification frame based on the Global Reporting Initiative Guideline and evaluated with the SOM. The SOM algorithms reduce multidimensional, large and complex data to lower dimensional maps which visualizes the distribution of the companies’ performance categories. The results indicate that the business cultures of the selected companies are similar. A closer examination reveals that the clustering by sector is more emphasized that by country. The more detailed determination of the SOM divides the paper sector further by their origin. Most obvious are the differences between sawmill and Finnish paper industry. The sawmill sector highlights especially product, service and macro-environment related aspects. Small German sawmills emphasized further tradition and history The paper sector stressed social responsibility, corporate structure, external activities and environmental issues. Inside the paper sector, the Finnish companies point out corporate strategy, development, organizational profile categories and economical performances. The German paper sector highlights in particular environmental issues and public relation. The web page analysis and related studies evaluate the corresponding categories similarly to the present research and thus support the results of the SOM approach.
  • Argyrou, Argyris (Svenska handelshögskolan, 2013)
    Economics and Society – 255
    The thesis examines how the auditing of journal entries can detect and prevent financial statement fraud. Financial statement fraud occurs when an intentional act causes financial statements to be materially misstated. Although it is not a new phenomenon, financial statement fraud has attracted much publicity in the wake of numerous cases of financial malfeasance (e.g. ENRON, WorldCom). Existing literature has provided limited empirical evidence on the link between auditing journal entries and financial statement fraud. The lack of evidence contrasts sharply with the responsibility of auditors to test the appropriateness of journal entries recorded in a general ledger. It becomes more pronounced when considering that journal entries pose a high risk of financial statement fraud, as the case of WorldCom has demonstrated. It is further exacerbated given that fraud results in considerable costs to a number of parties, for example: auditors may be exposed to litigation; investors may experience negative stock returns; and, capital markets may suffer from reduced liquidity. Motivated by these considerations, the thesis adopts the tenets of design-science research in order to develop three quantitative models for auditing journal entries. It first employs self-organizing map and extreme value theory to design the models as constructs. Subsequently, it codes the constructs in MATLAB to build functioning instantiations; and finally, it evaluates the instantiations by conducting a series of experiments on an accounting dataset containing journal entries. The contribution of the thesis lies in the proposed models and their potential applications in accounting. The first model can assist management to monitor the processing of journal entries as well as to assess the accuracy of financial statements. The second model can detect novel journal entries that differ from legitimate journal entries to such an extent that they could be ‘suspicious’. The third model can identify those journal entries that have both a very low probability of occurring and a monetary amount large enough to materially misstate financial statements. The thesis has a novelty value in that it investigates financial statement fraud from the unexplored perspective of journal entries. The thesis can lead to concrete practical applications in accounting, as the models can be implemented as a Computerised Assisted Audit Technique. This potentiality can be the focal point of additional research.
  • Junno, Niina; Koivisto, Emilia; Kukkonen, Ilmo; Malehmir, Alireza; Montonen, Markku (2019)
    We use self-organizing map (SOM) analysis to predict missing seismic velocity values from other available borehole data. The site of this study is the Kevitsa Ni-Cu-PGE deposit within the mafic-ultramafic Kevitsa intrusion in northern Finland. The site has been the target of extensive seismic reflection surveys, which have revealed a series of reflections beneath the Kevitsa resource area. The interpretation of these reflections has been complicated by disparate borehole data, particularly because of the scarce amount of available sonic borehole logs and the varying practices in logging of borehole lithologies. SOM is an unsupervised data mining method based on vector quantization. In this study, SOM is used to predict missing seismic velocities from other geophysical, geochemical, geological, and geotechnical data. For test boreholes, for which measured seismic velocity logs are also available, the correlation between actual measured and predicted velocities is strong to moderate, depending on the parameters included in the SOM analysis. Predicted reflectivity logs, based on measured densities and predicted velocities, show that some contacts between olivine pyroxenite/olivine websterite-dominant host rocks of the Kevitsa disseminated sulfide mineralization-and metaperidotite-earlier extensively used "lithology" label that essentially describes various degrees of alteration of different olivine pyroxenite variants-are reflective, and thus, alteration can potentially cause reflectivity within the Kevitsa intrusion.
  • Voutilainen, Ari; Arvola, Lauri Matti Juhani (2017)
    In order to improve our understanding of the connections between the biological processes and abiotic factors, we clustered complex long-term ecological data with the self-organizing map (SOM) technique. The available 21-year long (1990–2010) data set from a small pristine humic lake, in southern Finland, consisted of 27 meteorological, physical, chemical, and biological variables. The SOM grouped the data into three categories of which the first one was the largest with 12 variables, including metabolic processes, dissolved oxygen, total nitrogen and phosphorus, chlorophyll a, and taxonomical groups of plankton known to exist in spring. The second cluster comprised of water temperature and precipitation together with cyanobacteria, algae, rotifers, and crustacean zooplankton, an association emphasized with summer. The third cluster was consisted of six physical and chemical variables linked to autumn, and to the effects of inflow and/or water column mixing. SOM is a useful method for grouping the variables of such a large multi-dimensional data set, especially, when the purpose is to draw comprehensive conclusions rather than to search for associations across sporadic variables. Sampling should minimize the number of missing values. Even flexible statistical techniques, such as SOM, are vulnerable to biased results due to incomplete data.