Predicting diffusion of innovations with self-organisation and machine learning

The main subject of this master's thesis was predicting diffusion of innovations. The prediction was done in a special case: product has been available in some countries, and based on its diffusion in those countries the prediction is done for other countries. The prediction was based on finding similar countries with Self-Organizing Map~(SOM), using parameters of countries. Parameters included various economical and social key figures. SOM was optimised for different products using two different methods: (a) by adding diffusion information of products to the country parameters, and (b) by weighting the country parameters based on their importance for the diffusion of different products. A novel method using Differential Evolution (DE) was developed to solve the latter, highly non-linear optimisation problem. Results were fairly good. The prediction method seems to be on a solid theoretical foundation. The results based on country data were good. Instead, optimisation for different products did not generally offer clear benefit, but in some cases the improvement was clearly noticeable. The weights found for the parameters of the countries with the developed SOM optimisation method were interesting, and most of them could be explained by properties of the products.

Kylmämuovattujen putkipalkkien X-liitosten äärikestävyyden ja jäykkyyden määrittäminen sekä materiaalimallin kehittäminen

Työssä on tutkittu elementtimenetelmän avulla kylmämuovattujen nelikulmaisten putkipalkkien materiaalimallin kehittämistä ja putkipalkkien X-liitosten jäykkyyden ja äärikestävyyden määrittämistä. Työn tavoitteena on tutkia kylmämuovauksen vaikutuksia putkipalkkiprofiilin materiaaliominaisuuksiin materiaalikokeiden ja elementtianalyysien avulla sekä kehittää putkipalkille anisotrooppista materiaalimallia. Työssä määritettyjä materiaalimalleja on sovellettu X-liitosten elementtimalleihin, joiden käyttäytymistä on verrattu äärikestävyyskokeiden tuloksiin. Tutkimuksen perusteella Eurocode 3:n mitoitusohjeita voidaan turvallisesti soveltaa kylmämuovattujen putkipalkkien X-liitosten laskennassa. Työssä tehtyjen materiaalikokeiden ja elementtianalyysien perusteella materiaalin anisotrooppisuuden vaikutus liitoksen kestävyyteen on vähäistä, ja putkipalkin pituussuuntaista materiaalimallia voidaan soveltaa myös kehäsuuntaisille materiaaliominaisuuksille. Materiaalikokeiden simulointi osoittaa, että elementtimenetelmää voidaan käyttää materiaalimallin määrittämisen apuvälineenä.

Adaptive MCMC methods with applications in environmental and geophysical models

This work presents new, efficient Markov chain Monte Carlo (MCMC) simulation methods for statistical analysis in various modelling applications. When using MCMC methods, the model is simulated repeatedly to explore the probability distribution describing the uncertainties in model parameters and predictions. In adaptive MCMC methods based on the Metropolis-Hastings algorithm, the proposal distribution needed by the algorithm learns from the target distribution as the simulation proceeds. Adaptive MCMC methods have been subject of intensive research lately, as they open a way for essentially easier use of the methodology. The lack of user-friendly computer programs has been a main obstacle for wider acceptance of the methods. This work provides two new adaptive MCMC methods: DRAM and AARJ. The DRAM method has been built especially to work in high dimensional and non-linear problems. The AARJ method is an extension to DRAM for model selection problems, where the mathematical formulation of the model is uncertain and we want simultaneously to fit several different models to the same observations. The methods were developed while keeping in mind the needs of modelling applications typical in environmental sciences. The development work has been pursued while working with several application projects. The applications presented in this work are: a winter time oxygen concentration model for Lake Tuusulanjärvi and adaptive control of the aerator; a nutrition model for Lake Pyhäjärvi and lake management planning; validation of the algorithms of the GOMOS ozone remote sensing instrument on board the Envisat satellite of European Space Agency and the study of the effects of aerosol model selection on the GOMOS algorithm.

Lineaarisen hammashihnaservokäytön tilasäätö

Tuotantotehokkuus näyttelee yhä suurempaa roolia teollisuudessa, minkä vuoksi myös pakkauslinjas­toille joudutaan asettamaan suuria vaatimuksia. Usein leik­kaus- ja kappaleensiirtosovelluksissa käyte­tään lineaarisia ruuvikäyttöjä, jotka voitaisiin tietyin edellytyksin korvata halvemmilla ja osittain suori­tuskykyisimmillä hammashihnavetoisilla johteilla. Yleensä paikkasäädetty työsolu muodostuu kahden tai kolmen eri koordinaatisto­akselin suuntaan asen­netuista johteista. Tällaisen työsolun paikoitustarkkuuteen vaikuttavat muun muassa käytetty säätöra­kenne, moottorisäätöketjun viiveet, sekä laitteiston eri epälineaarisuudet, kuten kitka. Tässä työssä esitetään lineaarisen hammashihnaservokäytön dynaamista käytöstä kuvaava matemaatti­nen malli ja laaditaan mallin pohjalta laitteen simulointimalli. Mallin toimivuus varmistetaan käytän­nön identifiointitesteillä. Lisäksi työssä tut­kitaan, kuinka hyvään suorituskykyyn lineaarinen hammas­hihnaservokäyttö kyke­nee, jos teollisuudessa paikoitussäätörakenteena tyypillisesti käytetty kaskadira­kenne tai PID-rakenne korvataan kehittyneemmällä mallipohjaisella tilasäädinra­kenteella. Säädön toi­mintaa arvioidaan simulointien ja koelaitteistolla suoritetta­vien mittaus­ten perusteella.

Development of phase retrieval algorithm and techniques for optical spectrum analysis

Coherent anti-Stokes Raman scattering is the powerful method of laser spectroscopy in which significant successes are achieved. However, the non-linear nature of CARS complicates the analysis of the received spectra. The objective of this Thesis is to develop a new phase retrieval algorithm for CARS. It utilizes the maximum entropy method and the new wavelet approach for spectroscopic background correction of a phase function. The method was developed to be easily automated and used on a large number of spectra of different substances.. The algorithm was successfully tested on experimental data.

Multifactor Performance Study on Complex Systems for the Benefit of Management

The primary objective is to identify the critical factors that have a natural impact on the performance measurement system. It is important to make correct decisions related to measurement systems, which are based on the complex business environment. The performance measurement system is combined with a very complex non-linear factor. The Six Sigma methodology is seen as one potential approach at every organisational level. It will be linked to the performance and financial measurement as well as to the analytical thinking on which the viewpoint of management depends. The complex systems are connected to the customer relationship study. As the primary throughput can be seen in a new well-defined performance measurement structure that will also be facilitated as will an analytical multifactor system. These critical factors should also be seen as a business innovation opportunity at the same time. This master's thesis has been divided into two different theoretical parts. The empirical part consists of both action-oriented and constructive research approaches with an empirical case study. The secondary objective is to seek a competitive advantage factor with a new analytical tool and the Six Sigma thinking. Process and product capabilities will be linked to the contribution of complex system. These critical barriers will be identified by the performance measuring system. The secondary throughput can be recognised as the product and the process cost efficiencies which throughputs are achieved with an advantage of management. The performance measurement potential is related to the different productivity analysis. Productivity can be seen as one essential part of the competitive advantage factor.

Ultralujasta teräksestä valmistetun henkilönostimen teleskooppipuomin kehittäminen

Tässä työssä kehitettiin palo- ja pelastuskäyttöön tarkoitettuun henkilönostimeen teleskooppipuomin profiilit. Profiilien valmistusmateriaalina oli kuumavalssattu, ultraluja säänkestävä rakenneteräs. Työssä kehitettiin standardien ja ohjeiden pohjalta laskentapohja, jolla voidaan tutkia teleskooppipuomin jaksojen tukireaktioita, taivutus- ja vääntömomentteja ja leikkaus ja normaalivoimia. Laskentapohjassa voidaan varioida eri kuormitusten suuntia, teleskooppipuomin sivusuuntaista ulottumaa ja nostokulmaa. Profiilien alustavassa mitoituksessa hyödynnettiin paikallisen lommahduksen huomioon ottavia standardeja ja suunnitteluohjeita. Eri poikkileikkausten ominaisuuksia verrattiin keskenään ja profiili valittiin yhdessä kohdeyrityksen kanssa. Alustavan mitoituksen yhteydessä muodostettiin apuohjelma valitulle poikkileikkaukselle, jolla voitiin tutkia profiilin eri muuttujien vaikutusta mm. paikalliseen lommahdukseen ja jäykkyyteen. Laskentapohjaan sisällytettiin myös optimointirutiini, jolla voitiin minimoida poikkileikkauksen pinta-ala ja tätä kautta profiilin massa. Lopullinen mitoitus suoritettiin elementtimenetelmällä. Mitoituksessa tutkittiin alustavasti mitoitettujen profiilien paikallista lommahdusta lineaarisen stabiilius- ja epälineaarisen analyysin pohjalta. Profiilien jännityksiä tutkittiin tarkemmin mm. varioimalla kuormituksia ja osittelemalla elementtien normaalijännityksiä. Diplomityössä kehitetyllä ja analysoidulla teleskooppipuomilla voitiin keventää jaksojen painoja 15-30 %. Sivusuuntainen ulottuma parani samalla lähes 20 % ja nimelliskuorma kasvoi 25 %.

Learning Preferences with Kernel-Based Methods

Learning of preference relations has recently received significant attention in machine learning community. It is closely related to the classification and regression analysis and can be reduced to these tasks. However, preference learning involves prediction of ordering of the data points rather than prediction of a single numerical value as in case of regression or a class label as in case of classification. Therefore, studying preference relations within a separate framework facilitates not only better theoretical understanding of the problem, but also motivates development of the efficient algorithms for the task. Preference learning has many applications in domains such as information retrieval, bioinformatics, natural language processing, etc. For example, algorithms that learn to rank are frequently used in search engines for ordering documents retrieved by the query. Preference learning methods have been also applied to collaborative filtering problems for predicting individual customer choices from the vast amount of user generated feedback. In this thesis we propose several algorithms for learning preference relations. These algorithms stem from well founded and robust class of regularized least-squares methods and have many attractive computational properties. In order to improve the performance of our methods, we introduce several non-linear kernel functions. Thus, contribution of this thesis is twofold: kernel functions for structured data that are used to take advantage of various non-vectorial data representations and the preference learning algorithms that are suitable for different tasks, namely efficient learning of preference relations, learning with large amount of training data, and semi-supervised preference learning. Proposed kernel-based algorithms and kernels are applied to the parse ranking task in natural language processing, document ranking in information retrieval, and remote homology detection in bioinformatics domain. Training of kernel-based ranking algorithms can be infeasible when the size of the training set is large. This problem is addressed by proposing a preference learning algorithm whose computation complexity scales linearly with the number of training data points. We also introduce sparse approximation of the algorithm that can be efficiently trained with large amount of data. For situations when small amount of labeled data but a large amount of unlabeled data is available, we propose a co-regularized preference learning algorithm. To conclude, the methods presented in this thesis address not only the problem of the efficient training of the algorithms but also fast regularization parameter selection, multiple output prediction, and cross-validation. Furthermore, proposed algorithms lead to notably better performance in many preference learning tasks considered.

Reaction Kinetics and Viscosity Modelling in the Fusion Syntheses of Ca- and Ca/Mg-resinates

Rosin is a natural product from pine forests and it is used as a raw material in resinate syntheses. Resinates are polyvalent metal salts of rosin acids and especially Ca- and Ca/Mg- resinates find wide application in the printing ink industry. In this thesis, analytical methods were applied to increase general knowledge of resinate chemistry and the reaction kinetics was studied in order to model the non linear solution viscosity increase during resinate syntheses by the fusion method. Solution viscosity in toluene is an important quality factor for resinates to be used in printing inks. The concept of critical resinate concentration, c crit, was introduced to define an abrupt change in viscosity dependence on resinate concentration in the solution. The concept was then used to explain the non-inear solution viscosity increase during resinate syntheses. A semi empirical model with two estimated parameters was derived for the viscosity increase on the basis of apparent reaction kinetics. The model was used to control the viscosity and to predict the total reaction time of the resinate process. The kinetic data from the complex reaction media was obtained by acid value titration and by FTIR spectroscopic analyses using a conventional calibration method to measure the resinate concentration and the concentration of free rosin acids. A multivariate calibration method was successfully applied to make partial least square (PLS) models for monitoring acid value and solution viscosity in both mid-infrared (MIR) and near infrared (NIR) regions during the syntheses. The calibration models can be used for on line resinate process monitoring. In kinetic studies, two main reaction steps were observed during the syntheses. First a fast irreversible resination reaction occurs at 235 °C and then a slow thermal decarboxylation of rosin acids starts to take place at 265 °C. Rosin oil is formed during the decarboxylation reaction step causing significant mass loss as the rosin oil evaporates from the system while the viscosity increases to the target level. The mass balance of the syntheses was determined based on the resinate concentration increase during the decarboxylation reaction step. A mechanistic study of the decarboxylation reaction was based on the observation that resinate molecules are partly solvated by rosin acids during the syntheses. Different decarboxylation mechanisms were proposed for the free and solvating rosin acids. The deduced kinetic model supported the analytical data of the syntheses in a wide resinate concentration region, over a wide range of viscosity values and at different reaction temperatures. In addition, the application of the kinetic model to the modified resinate syntheses gave a good fit. A novel synthesis method with the addition of decarboxylated rosin (i.e. rosin oil) to the reaction mixture was introduced. The conversion of rosin acid to resinate was increased to the level necessary to obtain the target viscosity for the product at 235 °C. Due to a lower reaction temperature than in traditional fusion synthesis at 265 °C, thermal decarboxylation is avoided. As a consequence, the mass yield of the resinate syntheses can be increased from ca. 70% to almost 100% by recycling the added rosin oil.

Size Effect and Notch Size Effect in Metal Fatigue

It is a well known phenomenon that the constant amplitude fatigue limit of a large component is lower than the fatigue limit of a small specimen made of the same material. In notched components the opposite occurs: the fatigue limit defined as the maximum stress at the notch is higher than that achieved with smooth specimens. These two effects have been taken into account in most design handbooks with the help of experimental formulas or design curves. The basic idea of this study is that the size effect can mainly be explained by the statistical size effect. A component subjected to an alternating load can be assumed to form a sample of initiated cracks at the end of the crack initiation phase. The size of the sample depends on the size of the specimen in question. The main objective of this study is to develop a statistical model for the estimation of this kind of size effect. It was shown that the size of a sample of initiated cracks shall be based on the stressed surface area of the specimen. In case of varying stress distribution, an effective stress area must be calculated. It is based on the decreasing probability of equally sized initiated cracks at lower stress level. If the distribution function of the parent population of cracks is known, the distribution of the maximum crack size in a sample can be defined. This makes it possible to calculate an estimate of the largest expected crack in any sample size. The estimate of the fatigue limit can now be calculated with the help of the linear elastic fracture mechanics. In notched components another source of size effect has to be taken into account. If we think about two specimens which have similar shape, but the size is different, it can be seen that the stress gradient in the smaller specimen is steeper. If there is an initiated crack in both of them, the stress intensity factor at the crack in the larger specimen is higher. The second goal of this thesis is to create a calculation method for this factor which is called the geometric size effect. The proposed method for the calculation of the geometric size effect is also based on the use of the linear elastic fracture mechanics. It is possible to calculate an accurate value of the stress intensity factor in a non linear stress field using weight functions. The calculated stress intensity factor values at the initiated crack can be compared to the corresponding stress intensity factor due to constant stress. The notch size effect is calculated as the ratio of these stress intensity factors. The presented methods were tested against experimental results taken from three German doctoral works. Two candidates for the parent population of initiated cracks were found: the Weibull distribution and the log normal distribution. Both of them can be used successfully for the prediction of the statistical size effect for smooth specimens. In case of notched components the geometric size effect due to the stress gradient shall be combined with the statistical size effect. The proposed method gives good results as long as the notch in question is blunt enough. For very sharp notches, stress concentration factor about 5 or higher, the method does not give sufficient results. It was shown that the plastic portion of the strain becomes quite high at the root of this kind of notches. The use of the linear elastic fracture mechanics becomes therefore questionable.

Biotransformation and endocrine disruptive effects of contaminants in ringed seals - implications for monitoring and risk assessment

Marine mammals are exposed to persistent organic pollutants (POPs), which may be biotransformed to metabolites some of which are highly toxic. Both POPs and their metabolites may lead to adverse health effects, which have been studied using various biomarkers. Changes in endocrine homeostasis have been suggested to be sensitive biomarkers for contaminant-related effects. The overall objective of this doctoral thesis was to investigate biotransformation capacity of POPs and their potential endocrine disruptive effects in two contrasting ringed seal populations from the low contaminated Svalbard area and from the highly contaminated Baltic Sea. Biotransformation capacity was studied by determining the relationships between congener-specific patterns and concentrations of polychlorinated biphenyls (PCBs), organochlorine pesticides (OCPs), polybrominated diphenyl ethers (PBDEs) and their hydroxyl (OH)- and/or methylsulfonyl (MeSO₂)-metabolites, and catalytic activities of hepatic xenobiotic-metabolizing phase I and II enzymes. The results suggest that the biotransformation of PCBs, PBDEs and toxaphenes in ringed seals depends on the congener-specific halogen-substitution pattern. Biotransformation products detected in the seals included OH-PCBs, MeSO₂-PCBs and –DDE, pentachlorophenol, 4-OHheptachlorostyrene, and to a minor extent OH-PBDEs. The effects of life history state (moulting and fasting) on contaminant status and potential biomarkers for endocrine disruption, including hormone and vitamin homeostasis, were investigated in the low contaminated ringed seal population from Svalbard. Moulting/fasting status strongly affected thyroid, vitamin A and calcitriol homeostasis, body condition and concentrations of POPs and their OH-metabolites. In contrast, moulting/fasting status was not associated with variations in vitamin E levels. Endocrine disruptive effects on multiple endpoints were investigated in the two contrasting ringed seal populations. The results suggest that thyroid, vitamin A and calcitriol homeostasis may be affected by the exposure of contaminants and/or their metabolites in the Baltic ringed seals. Complex and non-linear relationships were observed between the contaminant levels and the endocrine variables. Positive relationships between circulating free and total thyroid hormone concentration ratios and OH-PCBs suggest that OH-PCBs may mediate the disruption of thyroid hormone transport in plasma. Species differences in thyroid and bone-related effects of contaminants were studied in ringed and grey seals from low contaminated references areas and from the highly contaminated Baltic Sea. The results indicate that these two species living at the same environment approximately at the same trophic level respond in a very different way to contaminant exposure. The results of this thesis suggest that the health status of the Baltic ringed seals has still improved during the last decade. PCB and DDE levels have decreased in these seals and the contaminant-related effects are different today than a decade ago. The health of the Baltic ringed seals is still suggested to be affected by the contaminant exposure. At the present level of the contaminant exposure the Baltic ringed seals seem to be at a zone where their body is able to compensate for the contaminant-mediated endocrine disruption. Based on the results of this thesis, several recommendations that could be applied on monitoring and assessing risk for contaminant effects are provided. Circulating OH-metabolites should be included in monitoring and risk assessment programs due to their high toxic potential. It should be noted that endogenous variables may have complex and highly variable responses to contaminant exposure including non-linear responses. These relationships may be further confounded by life history status. Therefore, it is highly recommended that when using variables related to endocrine homeostasis to investigate/monitor or assess the risk of contaminant effects in seals, the life history status of the animal should be carefully taken into consideration. This applies especially when using thyroid, vitamin A or calcitriolrelated parameters during moulting/fasting period. Extrapolations between species for assessing risk for contaminant effects in phocid seals should be avoided.

Analysis of patterns in mining time series using Qlucore

Raw measurement data does not always immediately convey useful information, but applying mathematical statistical analysis tools into measurement data can improve the situation. Data analysis can offer benefits like acquiring meaningful insight from the dataset, basing critical decisions on the findings, and ruling out human bias through proper statistical treatment. In this thesis we analyze data from an industrial mineral processing plant with the aim of studying the possibility of forecasting the quality of the final product, given by one variable, with a model based on the other variables. For the study mathematical tools like Qlucore Omics Explorer (QOE) and Sparse Bayesian regression (SB) are used. Later on, linear regression is used to build a model based on a subset of variables that seem to have most significant weights in the SB model. The results obtained from QOE show that the variable representing the desired final product does not correlate with other variables. For SB and linear regression, the results show that both SB and linear regression models built on 1-day averaged data seriously underestimate the variance of true data, whereas the two models built on 1-month averaged data are reliable and able to explain a larger proportion of variability in the available data, making them suitable for prediction purposes. However, it is concluded that no single model can fit well the whole available dataset and therefore, it is proposed for future work to make piecewise non linear regression models if the same available dataset is used, or the plant to provide another dataset that should be collected in a more systematic fashion than the present data for further analysis.

Learning Robot Motions from Visual Sensing

Learning from demonstration becomes increasingly popular as an efficient way of robot programming. Not only a scientific interest acts as an inspiration in this case but also the possibility of producing the machines that would find application in different areas of life: robots helping with daily routine at home, high performance automata in industries or friendly toys for children. One way to teach a robot to fulfill complex tasks is to start with simple training exercises, combining them to form more difficult behavior. The objective of the Master’s thesis work was to study robot programming with visual input. Dynamic movement primitives (DMPs) were chosen as a tool for motion learning and generation. Assuming a movement to be a spring system influenced by an external force, making this system move, DMPs represent the motion as a set of non-linear differential equations. During the experiments the properties of DMP, such as temporal and spacial invariance, were examined. The effect of the DMP parameters, including spring coefficient, damping factor, temporal scaling, on the trajectory generated were studied.

Harustetun 110 kV kannatuspylvään konseptin kehittäminen

Diplomityössä kehitettiin harustetun 110 kV kannatuspylvään konsepti tuotteeksi. Pylväs on säänkestävästä teräksestä valmistettu putkipalkkirakenteinen I-pylväs. Tavoitteena oli suunnitella rakenteesta kokonaistaloudellisesti edullinen. Rakenteen suunnittelussa otettiin huomioon valmistus-, kuljetus- ja varastointi- sekä rakentamisnäkökohtia. Työssä perehdyttiin pylväsrakenteiden yksityiskohtiin, putkipalkkien liitosmenetelmiin ja pylvään jalan nivelöintiratkaisuihin. Säänkestävä rakennemateriaali otettiin huomioon rakennesuunnittelussa. Rakenteen lujuusteknisen suunnittelun apuna käytettiin epälineaarista elementtimenetelmää. Pylväsrakenteen käyttäytyminen mallinnettiin geometrisesti epälineaariseksi, ja liitosdetaljien analysointia varten kehitettiin epälineaarisia materiaalimalleja. Rakenteen värähtelykäyttäytyminen analysoitiin myös elementtimenetelmällä. Lopputuloksena saatiin aikaan pylväs, joka täyttää sille asetetut vaatimukset. Pylväs on helposti valmistettava, kuljetettava ja pystytettävä.

On convergence of transforms based on parabolic scaling

This thesis studies properties of transforms based on parabolic scaling, like Curvelet-, Contourlet-, Shearlet- and Hart-Smith-transform. Essentially, two di erent questions are considered: How these transforms can characterize H older regularity and how non-linear approximation of a piecewise smooth function converges. In study of Hölder regularities, several theorems that relate regularity of a function f : R2 → R to decay properties of its transform are presented. Of particular interest is the case where a function has lower regularity along some line segment than elsewhere. Theorems that give estimates for direction and location of this line, and regularity of the function are presented. Numerical demonstrations suggest also that similar theorems would hold for more general shape of segment of low regularity. Theorems related to uniform and pointwise Hölder regularity are presented as well. Although none of the theorems presented give full characterization of regularity, the su cient and necessary conditions are very similar. Another theme of the thesis is the study of convergence of non-linear M ─term approximation of functions that have discontinuous on some curves and otherwise are smooth. With particular smoothness assumptions, it is well known that squared L2 approximation error is O(M-2(logM)3) for curvelet, shearlet or contourlet bases. Here it is shown that assuming higher smoothness properties, the log-factor can be removed, even if the function still is discontinuous.