The main purpose of this degree project is to reveal the Airbnb customer’s preferences and quantify the impact of non-market factors on the market price of tourist accommodation in Berlin, Germany. The data retrieved from Airbnb listings, publicly available on Inside Airbnb (2017), was supplemented on indicator of sharing economy accommodation using machine learning method in order to distinguish between amateur and business-running professional hosts. The main aim is to examine the consumers’ preferences and quantify the marginal effect of "real sharing economy" accommodation and other key variables on market price. This is accomplished by model approach using hedonic pricing method, which is used to estimate the economic value of particular attribute. Surprisingly, our data indicates the negative impact of sharing economy indicator on price. The set of motivations of consumers, which determine their valuation of Airbnb listings, was identified. The trade-off between encompass and parsimony of the set was desired in order to build an effective model. Calculation of proportion of explained variance showed that the price is affected mainly by number of accommodated persons, degree of privacy, number of bedrooms, cancellation policy, distance from the city centre and sharing economy indicator in decreasing order.

Orientador: Prof. Dr. Fabiano Fragoso Costa / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia Elétrica, 2014. / Este trabalho propõe um algoritmo de estimaçãoo fasorial baseado na versão modificada do algoritmo de mínimos quadrados recursivo. Este algoritmo é adequado para proteção de sistemas de potência, uma vez que sua resposta é rapida e robusta 'a presença da componente dc de decaimento exponencial, que é uma interferência comum em condi¸ções de falta atrasando a convergência da estimativa fasorial. Além disso, esta dissertação também investiga o uso do chamado método de Prony, a fim de auxiliar e acelerar a convergência da estimação fasorial do algoritmo dos mínimos quadrados. O método de Prony determina o decaimento exponencial a ser extraído do sinal analisado. As técnicas desenvolvidas nessa disserta¸c¿ao foram comparadas com o tradicional estimador de Fourier de um ciclo atrav'es de simula¸c¿oes realizadas em Matlab e de experimentos realizados com um processador de sinais e um amplificador de sinais. Os resultados mostram melhorias da t'ecnica proposta em comparação ao algoritmo de Fourier e incentivam futuras pesquisas relacionadas a este assunto. / This work proposes a phasor estimation algorithm based on a modified recursive least-squares. This algorithm is suitable for power systems protection once its response quick and robust to the decaying dc component, which is a most usual interference in fault conditions and delays the phasor estimation convergence. Furthermore, this dissertation also investigates the usage of the so-called Prony¿s method in order to aid and to speed up the least-squares phasor estimation convergence. This method determines the exponential decaying to be extracted out of the analyzed signal. The present developed techniques have been compared with the traditional one-cycle Fourier estimation by simulation performed on Matlab and by experiments accomplished with a digital signal processor and a signal amplifier. The results show improvements of the proposed techniques over the Fourier algorithm and encourage further research in this topic.

A produção agrícola tem crescido nos últimos anos impulsionada pelo aumento populacional e por avanços tecnológicos. Esse crescimento pode ocasionar impactos ambientais importantes, inclusive a degradação do solo, se não forem realizados o correto planejamento agrícola e manejo do solo, a fim de assegurar uma produção competitiva e sustentável. Para isto, a descrição da variabilidade espacial do solo é necessária, sendo realizada convencionalmente através de coleta e análise de amostras. Entretanto, estes métodos convencionais de levantamento da variabilidade do solo possuem custos elevados e demandam bastante tempo e mãode- obra para serem realizados. Com o aumento da quantidade de informação necessária os custos para descrição da variabilidade espacial do solo podem tornarse um obstáculo, se somente metodologias convencionais são aplicadas. Portanto, métodos alternativos tornam-se necessários para auxiliar no levantamento de atributos do solo em escala adequada ao manejo agrícola. Para isto, são propostas metodologias de espectroscopia de reflectância no Vis-NIR-SWIR, as quais empregam o comportamento espectral do solo de 400 nm a 2500 nm para realizar a quantificação de seus atributos. Isto é possível, pois a informação espectral possui relação direta com os constituintes do solo. Assim, no presente estudo é avaliado o uso de imagens aéreas hiperespectrais na quantificação de atributos do solo, através do método PLSR, e mapeamento destes atributos, empregando krigagem. O desempenho das predições feitas com dados do sensor aéreo é comparado ao obtido com espectros coletados em laboratório. Foi também avaliado, através de experimentos de campo, com diferentes doses de calcário em duas áreas diferentes (textura arenosa e média), o uso de informações espectrais coletadas no campo, em movimento, e em laboratório, com amostras úmidas, para a quantificação de atributos e da necessidade de calagem, pelo método PLSR. Foram obtidos resultados satisfatórios através dos dados do sensor aéreo, principalmente, para a quantificação dos teores de argila, areia e CTC (R2 de 0,73, 0,73 e 0,80, respectivamente). Com relação aos dados obtidos por sensoriamento próximo no campo, os melhores resultados foram obtidos para a área de textura média, com R2 de 0,33, 0,38 e 0,61 para a predição da CTC, V% e da necessidade de calagem, respectivamente. / Agricultural production has increased in the last years stimulated by the population growth and technological advances. This can cause significant environmental impacts including soil degradation if suitable agricultural planning and soil management are not applied in order to ensure a competitive and sustainable production. For this purpose, the soil variability assessment is needed and it is conventionally performed through soil sampling and analysis. However, conventional methods have high costs and require considerable time and labor. When the amount of information needed increases, costs to describe soil spatial variability may become an obstacle if only conventional methodologies are applied. Therefore, alternative methods can help to depict soil properties variability on a scale suitable to soil management. So, Vis-NIR-SWIR reflectance spectroscopy (from 400 nm to 2500 nm) is proposed as a mean to predict soil properties. This is possible because spectral information has a direct relationship with soil constituents and characteristics. Therefore, in this study hyperspectral airborne imagery is evaluated as an information source to be used in soil properties quantification, via the PLSR method, and mapping, using kriging. The performance of the models derived from airborne imagery data was compared with the results obtained by models calculated from laboratory sensor data. The use of spectral information collected in the field (on-thego) was evaluated too using a field experiment in witch different rates of lime were applied. The experiment was allocated in two fields with different soil textures (one with about 100 g kg-1 of clay and other with about 320 g kg-1 of clay). The soil properties prediction based on the on-the-go spectral measurements were compared to predictions made using spectra collect in the laboratory and the PLSR method was used to calculate models. Satisfactory results were obtained with airborne sensor data, especially for clay, sand and CTC quantification (R2 of 0.73, 0.73 and 0.80, respectively). Regarding the on-the-go proximal sensing, better predictions were obtained for the clayey area, with R2 of 0.33, 0.38 and 0.61 for predictions of CEC, base saturation of the soil CEC (V%) and lime requirement, respectively.

We study nonlinear least-squares problem that can be transformed to linear problem by change of variables. We derive a general formula for the statistically optimal weights and prove that the resulting linear regression gives an optimal estimate (which satisfies an analogue of the Rao–Cramer lower bound) in the limit of small noise.

No description available.

The GIS Industry has digitised cadastre from the best available paper maps over the last few decades, incorporating the inherent errors in those paper maps and in the digitising process. The advent of Global Positioning Systems, modern surveying instruments and advances in the computing industry has made it desirable and affordable to upgrade the placement, in terms of absolute and relative position) of these digital cadastres. The Utility Industry has used GIS software to place their assets relative to these digital cadastres, and are now finding their assets placed incorrectly when viewed against these upgraded digital cadastres. This thesis examines the processes developed in the software program called the ???Spatial Adjustment Engine???, and documents a holistic approach to semi-automating the upgrading of the digital cadastre and the subsequent upgrading of the utility assets. This thesis also documents the various pilot projects undertaken during the development of the Spatial Adjustment Engine, the topological scenarios found in each pilot, their solution, and provides a framework of definitions needed to explore this field further. The results of each pilot project are given in context, and lead to the conclusions. The conclusions indicate the processes and procedures implemented in the Spatial Adjustment Engine are a suitable mechanism for the upgrade of digital cadastre and of spatially dependant themes such as utility assets, zoning themes, annotation layers, and some road centreline themes.

This thesis describes the research undertaken for the degree of Doctor of Philosophy, testing the hypothesis that spectrometer data can be used to establish usable relationships for prediction of pasture quality attributes. The research data consisted of reflectance measurements of various temperate pasture types recorded at four different times (years 2000 to 2002), recorded by three hyperspectral sensors, the in situ ASD, the airborne HyMap and the satellite-borne Hyperion. Corresponding ground-based pasture samples were analysed for content of chlorophyll, water, crude protein, digestibility, lignin and cellulose at three study sites in rural Victoria, Australia. This context was used to evaluate effects of sensor differences, data processing and enhancement, analytical methods and sample variability on the predictive capacity of derived prediction models. Although hyperspectral data analysis is being applied in many areas very few studies on temperate pastures have been conducted and hardly any encompass the variability and heterogeneity of these southern Australian examples. The research into the relationship between the spectrometer data and pasture quality attribute assays was designed using knowledge gained from assessment of other hyperspectral remote sensing and near-infrared spectroscopy research, including bio-chemical and physical properties of pastures, as well as practical issues of the grazing industries and carbon cycling/modelling. Processing and enhancement of the spectral data followed methods used by other hyperspectral researchers with modifications deemed essential to produce better relationships with pasture assay data. As many different methods are in use for the analysis of hyperspectral data several alternative approaches were investigated and evaluated to determine reliability, robustness and suitability for retrieval of temperate pasture quality attributes. The analyses employed included stepwise multiple linear regression (SMLR) and partial least squares regression (PLSR). The research showed that the spectral research data had a higher potential to be used for prediction of crude protein and digestibility than for the plant fibres lignin and cellulose. Spectral transformation such as continuum removal and derivatives enhanced the results. By using a modified approach based on sample subsets identified by a matrix of subjective bio-physical and ancillary data parameters, the performance of the models were enhanced. Prediction models from PLSR developed on ASD in situ spectral data, HyMap airborne imagery and Hyperion and corresponding pasture assays showed potential for predicting the two important pasture quality attributes crude protein and digestibility in hyperspectral imagery at a few quantised levels corresponding to levels currently used in commercial feed testing. It was concluded that imaging spectrometry has potential to offer synoptic, simultaneous and spatially continuous information valuable to feed based enterprises in temperate Victoria. The thesis provide a significant contribution to the field of hyperspectral remote sensing and good guidance for future hyperspectral researchers embarking on similar tasks. As the research is based on temperate pastures in Victoria, Australia, which are dominated by northern hemisphere species, the findings should be applicable to analysis of temperate pastures elsewhere, for example in Western Australia, New Zealand, South Africa, North America, Europe and northern Asia (China).

This thesis proposes, analyses and implements a fast and accurate real-time power system disturbances identification method based on an enhanced linear least squares algorithm for mitigation and monitoring of various power quality problems such as current harmonics, grid unbalances and voltage dips. The enhanced algorithm imposes less real-time computational burden on processing the system and is thus called ???efficient least squares algorithm???. The proposed efficient least squares algorithm does not require matrix inversion operation and contains only real numbers. The number of required real-time matrix multiplications is also reduced in the proposed method by pre-performing some of the matrix multiplications to form a constant matrix. The proposed efficient least squares algorithm extracts instantaneous sine and cosine terms of the fundamental and harmonic components by simply multiplying a set of sampled input data by the pre-calculated constant matrix. A power signal processing system based on the proposed efficient least squares algorithm is presented in this thesis. This power signal processing system derives various power system quantities that are used for real-time monitoring and disturbance mitigation. These power system quantities include constituent components, symmetrical components and various power measurements. The properties of the proposed power signal processing system was studied using modelling and practical implementation in a digital signal processor. These studies demonstrated that the proposed method is capable of extracting time varying power system quantities quickly and accurately. The dynamic response time of the proposed method was less than half that of a fundamental cycle. Moreover, the proposed method showed less sensitivity to noise pollution and small variations in fundamental frequency. The performance of the proposed power signal processing system was compared to that of the popular DFT/FFT methods using computer simulations. The simulation results confirmed the superior performance of the proposed method under both transient and steady-state conditions. In order to investigate the practicability of the method, the proposed power signal processing system was applied to two real-life disturbance mitigation applications namely, an active power filter (APF) and a distribution synchronous static compensator (D-STATCOM). The validity and performance of the proposed signal processing system in both disturbance mitigations applications were investigated by simulation and experimental studies. The extensive modelling and experimental studies confirmed that the proposed signal processing system can be used for practical real-time applications which require fast disturbance identification such as mitigation control and power quality monitoring of power systems

The objective of this research is to empirically assess the value and validity of a multi-organization data set in the building of prediction models for several ‘local’ software organizations; that is, smaller organizations that might have a few project records but that are interested in improving their ability to accurately predict software project effort. Evidence to date in the research literature is mixed, due not to problems with the underlying research ideas but with limitations in the analytical processes employed: • the majority of previous studies have used only a single organization as the ‘local’ sample, introducing the potential for bias • the degree to which the conclusions of these studies might apply more generally is unable to be determined because of a lack of transparency in the data analysis processes used. It is the aim of this research to provide a more robust and visible test of the utility of the largest multi-organization data set currently available – that from the ISBSG – in terms of enabling smaller-scale organizations to build relevant and accurate models for project-level effort prediction. Stepwise regression is employed to enable the construction of ‘local’, ‘global’ and ‘refined global’ models of effort that are then validated against actual project data from eight organizations. The results indicate that local data, that is, data collected for a single organization, is almost always more effective as a basis for the construction of a predictive model than data sourced from a global repository. That said, the accuracy of the models produced from the global data set, while worse than that achieved with local data, may be sufficiently accurate in the absence of reliable local data – an issue that could be investigated in future research. The study concludes with recommendations for both software engineering practice – in setting out a more dynamic scenario for the management of software development – and research – in terms of implications for the collection and analysis of software engineering data.

<p>The paper describes predictive modelling of biomarker data stemming from patients suffering from multiple sclerosis. Improvements of multivariate analyses of the data are investigated with the goal of increasing the capability to assign samples to correct subgroups from the data alone.</p><p>The effects of different preceding scalings of the data are investigated and combinations of multivariate modelling methods and variable selection methods are evaluated. Attempts at merging the predictive capabilities of the method combinations through voting-procedures are made. A technique for improving the result of PLS-modelling, called bagging, is evaluated.</p><p>The best methods of multivariate analysis of the ones tried are found to be Partial least squares (PLS) and Support vector machines (SVM). It is concluded that the scaling have little effect on the prediction performance for most methods. The method combinations have interesting properties – the default variable selections of the multivariate methods are not always the best. Bagging improves performance, but at a high cost. No reasons for drastically changing the work flows of the biomarker data analysis are found, but slight improvements are possible. Further research is needed.</p>