Browsing by Subject "statistical analysis"

Sort by: Order: Results:

Now showing items 1-8 of 8
  • Väisänen, Rauno; Heliövaara, Kari (The Society of Forestry in Finland - The Finnish Forest Research Institute, 1994)
    The presence/absence data of twenty-seven forest insect taxa (e.g. Retinia resinella, Formica spp., Pissodes spp., several scolytids) and recorded environmental variation were used to investigate the applicability of modelling insect occurrence based on satellite imagery. The sampling was based on 1800 sample plots (25 m by 25 m) placed along the sides of 30 equilateral triangles (side 1 km) in a fragmented forest area (approximately 100 km2) in Evo, S Finland. The triangles were overlaid on land use maps interpreted from satellite images (Landsat TM 30 m multispectral scanner imagery 1991) and digitized geological maps. Insect occurrence was explained using either environmental variables measured in the field or those interpreted from the land use and geological maps. The fit of logistic regression models varied between species, possibly because some species may be associated with the characteristics of single trees while other species with stand characteristics. The occurrence of certain insect species at least, especially those associated with Scots pine, could be relatively accurately assessed indirectly on the basis of satellite imagery and geological maps. Models based on both remotely sensed and geological data better predicted the distribution of forest insects except in the case of Xylechinus pilosus, Dryocoetes sp. and Trypodendron lineatum, where the differences were relatively small in favour of the models based on field measurements. The number of species was related to habitat compartment size and distance from the habitat edge calculated from the land use maps, but logistic regressions suggested that other environmental variables in general masked the effect of these variables in species occurrence at the present scale.
  • Verkasalo, Erkki; Mottonen, Veikko; Roitto, Marja; Vepsalainen, Jouko; Kumar, Anuj; Ilvesniemi, Hannu; Siwale, Workson; Julkunen-Tiitto, Riitta; Raatikainen, Olavi; Sikanen, Lauri (2021)
    This study aimed to identify and quantify phenolic and resin acid extractive compounds in Scots pine stemwood and sawmill residues in four climatic regions of Finland to evaluate their most optimal sources for bio-based chemical biorefining and bioenergy products. The sample consisted of 140 trees from 28 stands, and sawdust lots from 11 log stands. NMR for the overall extractive analysis and HPLC for the quantitative estimation of phenolic and resin acid compounds were employed. Correlation analysis, multivariate factor analysis, principle component analysis and multiple linear regression modelling were applied for statistical analysis. HPLC identified 12 extractive compounds and NMR five more resin acids. Pinosylvin (PS), pinosylvin monomethyl ether (PSMME), and partly neolignans/lignans occurred in the largest concentrations. Wood type caused the most variation, heartwood having larger concentrations than sapwood (sawdust between them). Regional differences in the concentrations were smaller, but factor analysis distinguished the northern and the southern regions into their own groups. The results indicated higher concentrations of PS, PSMME, and vanillic acid in southern regions and those of, e.g., PSMME glycoside, lignan 2, and neolignan 1 in northern regions. The rather low concentrations of extractives in stemwood and sawdust imply value-added products, efficient sorting and/or large raw material volumes.
  • Toukola, Peppi (Helsingin yliopisto, 2021)
    In this thesis the suitability of Nuclear Magnetic Resonance (NMR) spectroscopy in the identification of rubbers in museum collections is discussed through a literature review and experimental work where samples from the rubber collection of Tampere Museums were analysed with different NMR techniques. The literature part of this thesis focuses on recent (2011-2020) scientific publications on analytical instrumental techniques used in the identification of cultural heritage plastics. Vibrational spectroscopy methods utilizing hand-held or portable devices have been the most prominent methods used in characterization of historical plastics materials. Bench-top devices and analytical techniques requiring sampling were used to acquire more detailed analysis results. However, NMR spectroscopy was not used as the main analysis technique in the reviewed publications. In the experimental part altogether 21 rubber object samples and 8 reference samples were identified using 1D and 2D NMR techniques in solution state. Three samples were additionally analysed with solid-state High Resolution Magic Angle Spinning (HRMAS) NMR spectroscopy. The chemical structures of the samples were confirmed with these methods. To further explore fast and more automated identification of the rubber samples a statistical classification model utilizing acquired solution-state 1H NMR data was developed. Three rubber types were chosen for the analysis. The model was created using analysis data from the museum object samples and validated using the reference sample data. Identification rate of 100 % was achieved.
  • Pesonen, Mauno; Kettunen, Arto; Räsänen, Petri (The Society of Forestry in Finland - The Finnish Forest Research Institute, 1995)
    The factors affecting the non-industrial, private forest landowners' (hereafter referred to using the acronym NIPF) strategic decisions in management planning are studied. A genetic algorithm is used to induce a set of rules predicting potential cut of the landowners' choices of preferred timber management strategies. The rules are based on variables describing the characteristics of the landowners and their forest holdings. The predictive ability of a genetic algorithm is compared to linear regression analysis using identical data sets. The data are cross-validated seven times applying both genetic algorithm and regression analyses in order to examine the data-sensitivity and robustness of the generated models. The optimal rule set derived from genetic algorithm analyses included the following variables: mean initial volume, landowner's positive price expectations for the next eight years, landowner being classified as farmer, and preference for the recreational use of forest property. When tested with previously unseen test data, the optimal rule set resulted in a relative root mean square error of 0.40. In the regression analyses, the optimal regression equation consisted of the following variables: mean initial volume, proportion of forestry income, intention to cut extensively in future, and positive price expectations for the next two years. The R2 of the optimal regression equation was 0.34 and the relative root mean square error obtained from the test data was 0.38. In both models, mean initial volume and positive stumpage price expectations were entered as significant predictors of potential cut of preferred timber management strategy. When tested with the complete data set of 201 observations, both the optimal rule set and the optimal regression model achieved the same level of accuracy.
  • Avela, Henri (Helsingin yliopisto, 2019)
    Lipidomics is a quickly growing trend in metabolomics research: not only seen as passive cell membrane building blocks, lipids contribute actively to cell signaling and identification, thus seen as potential biomarkers (e.g. for early stage cancer diagnostics). The literature part includes a review of 63 articles on UHPLC/MS-methods in the time frame of 2017-05/2019. The following literature is focused especially on glycerophospholipids (GPs). In addition, an overview to basic glycerolipids (GLs) and sphingolipids (SPs) is established, which evidently affects the emphasis and narration of lipid class representations in this review. Chromatographic methods in lipidomics are used to achieve either very selective or all-encompassing analyses for lipid classes. Since HPLC/MS is an insufficient method for fully encompassing low-abundance lipids, UHPLC/MS was mostly used for metabolic profiling where its large analyte range due to high sensitivity, separation efficiency and resolution excels in performance compared to other methods. Imaging techniques have further diverted towards DIMS and other novel non-chromatographic methods, e.g. Raman techniques with single cell resolution. The field of mass-spectral lipidomics is divided between studies using isotope-labeled standards or fully standardless algorithm-based analyses, furthermore, machine learning and statistical analysis has increased. The experimental part focused on LC-IMS-MS and plasma-based in-house database method development for targeted analysis of ascites. Method development included optimization of the chromatography, adduct species selection and data-independent/-dependent fragmentation. Totally, 130 potential species from the LIPID MAPS database were used for the identification at the minimum score of 79% for identification in the Qualitative Workflows with retention times (RTs) and Mass Profiler-program with collision cross-sections (CCSs). Plasma sample analyses resulted in the documentation of 70 RTs and 36 CCS values. Two lipid extraction methods (Folch and BUME) with pre-sampling surrogates and post-sampling internal standards were compared with each other. The process resulted in confirming the BUME method in lipidomics to be superior in ecology-, workload-, health- and extraction-related properties. The lipidome of ascites has rarely been studied due to its availability only in diseased patients. Also, limiting factors for these studies are the logistics to realise such a representative analysis.
  • Kilkki, Pekka (Suomen metsätieteellinen seura, 1983)
  • Hannula, Jukka (2001)
    The objective of the thesis was to utilize the measured temperature- and relative humidity data in order to determine the usage environment of mobile phones. New statistical method for lifetime prediction of solder joint is represented in this study. Two fundamental theories of statistics that are the basis for the method are also represented. The data were also utilized for three physical lifetime estimating models. The measured temperature data were parametrized for the lifetime estimations. The effects of three usage environments to the lifetime of solder joint were studied by using represented models and the results were compared. The average values and standard deviations were calculated from the temperature- and relative humidity data. The Measured data were also presented in 3D-figures of the temperature and relative humidity. Average values of temperature varied between 22 – 30 ˚C and values of relative humidity between 30 – 50 %. The measured usage environments were also divided into meteorological regions, and global coverage of measurements was also discussed. The measured temperature data was simplified by linearization and parameters for lifetime estimations were calculated. Four moments of the original and the modified data were calculated and compared in order to check the representativeness of the modified data. The information of the modified data corresponded with the original data, and therefore, it could be utilized in the lifetime estimation models. The severity of usage environments was studied by estimating the lifetime of the solder joint. The fatigue of the solder joint caused by climatic temperature variation was observed to be the most severe in the low temperature environment. Severity of the temperature cycling environment was higher than than that of the high temperature environment, which was observed to be the least severe environment. Results from this study can be utilized in developing of testing- and design guidelines, as a stress factor in simulation models, in usage environment profiling and in international standardization work.
  • Kilkki, Pekka; Varmola, Martti (Suomen metsätieteellinen seura, 1981)