Efficient price management is crucial for companies with many different products to keep track of, leading to the common practice of price logging. Today, these prices are often adjusted manually, but setting prices manually can be labor-intensive and prone to human error. This project aims to use machine learning to assist in the pricing of products by estimating the prices to be inserted. Multiple machine learning models have been tested, and an artificial neural network has been implemented for estimating prices effectively. Through additional experimentation, the design of the network was fine-tuned to make it compatible with the project’s needs. The libraries used for implementing and managing the machine learning models are mainly ScikitLearn and TensorFlow. As a result, the trained model has been saved into a file and integrated with an API for accessibility.

Text in Afrikaans / The purpose of the research is to compare the quantity and quality of reporting on the employees of companies with the profitability and size of the top 100 companies over a four year period. The research is limited to the availability of quality information on employees in the financial statements of companies. A literature study is followed by an emperical inquiry into a comparison of the reporting on employees with the profitability and size of companies. The results reflect the following: * the quantity and quality of reporting on the employees of companies does not inevitably show an increase; * a significant relationship exist between the size and profitability of companies; * the role of reporting on the employees of companies is significant with regards to profit determination; and * some of the companies performed uniquely with regards to the quantity and quality of reporting on the employees of companies. / Accounting / M.Comm.(Accounting)

Darbas apima banko klientų kredito istorinių duomenų dėsningumų tyrimą. Pirmiausia nagrinėjamos banko duomenų saugyklos, siekiant kuo geriau perprasti bankinius duomenis. Vėliau naudojant banko duomenų imtis, kurios apima kreditų grąžinimo istoriją, siekiama įvertinti klientų nemokumo riziką. Tai atliekama adaptuojant algoritmus bei programinę įrangą duomenų tyrimui, kuris pradedamas nuo informacijos apdorojimo ir paruošimo. Paskui pritaikant įvairius klasifikavimo algoritmus, sudarinėjami modeliai, kuriais siekiama kuo tiksliau suskirstyti turimus duomenis, nustatant nemokius klientus. Taip pat siekiant įvertinti kliento vėluojamų mokėti paskolą dienų skaičių pasitelkiami regresijos algoritmai bei sudarinėjami prognozės modeliai. Taigi darbo metu atlikus numatytus tyrimus, pateikiami duomenų vitrinų modeliai, informacijos srautų schema. Taip pat nurodomi klasifikavimo ir prognozavimo modeliai bei algoritmai, geriausiai įvertinantys duotas duomenų imtis. / This work is about analysing regularities in bank clients historical credit data. So first of all bank information repositories are analyzed to comprehend banks data. Then using data mining algorithms and software for bank data sets, which describes credit repayment history, clients insolvency risk is being tried to estimate. So first step in analyzis is information preprocessing for data mining. Later various classification algorithms is used to make models wich classify our data sets and help to identify insolvent clients as accurate as possible. Besides clasiffication, regression algorithms are analyzed and prediction models are created. These models help to estimate how long client are late to pay deposit. So when researches have been done data marts and data flow schema are presented. Also classification and regressions algorithms and models, which shows best estimation results for our data sets, are introduced.

The use of audio and video communication has increased exponentially over the last decade and has gone from speech over GSM to HD resolution video conference between continents on mobile devices. As the use becomes more widespread the interest in delivering high quality media increases even on devices with limited resources. This includes both development and enhancement of the communication chain but also the topic of objective measurements of the perceived quality. The focus of this thesis work has been to perform enhancement within speech encoding and video decoding, to measure influence factors of audio and video performance, and to build methods to predict the perceived video quality. The audio enhancement part of this thesis addresses the well known problem in the GSM system with an interfering signal generated by the switching nature of TDMA cellular telephony. Two different solutions are given to suppress such interference internally in the mobile handset. The first method involves the use of subtractive noise cancellation employing correlators, the second uses a structure of IIR notch filters. Both solutions use control algorithms based on the state of the communication between the mobile handset and the base station. The video enhancement part presents two post-filters. These two filters are designed to improve visual quality of highly compressed video streams from standard, block-based video codecs by combating both blocking and ringing artifacts. The second post-filter also performs sharpening. The third part addresses the problem of measuring audio and video delay as well as skewness between these, also known as synchronization. This method is a black box technique which enables it to be applied on any audiovisual application, proprietary as well as open standards, and can be run on any platform and over any network connectivity. The last part addresses no-reference (NR) bitstream video quality prediction using features extracted from the coded video stream. Several methods have been used and evaluated: Multiple Linear Regression (MLR), Artificial Neural Network (ANN), and Least Square Support Vector Machines (LS-SVM), showing high correlation with both MOS and objective video assessment methods as PSNR and PEVQ. The impact from temporal, spatial and quantization variations on perceptual video quality has also been addressed, together with the trade off between these, and for this purpose a set of locally conducted subjective experiments were performed.

Artificial Neural Networks (ANNs) have been comprehensively researched, both from a computer scientific perspective and with regard to their use for predictive modelling in a wide variety of applications including hydrology and the environment. Yet their adoption for live, real-time systems remains on the whole sporadic and experimental. A plausible hypothesis is that this may be at least in part due to their treatment heretofore as “black boxes” that implicitly contain something that is unknown, or even unknowable. It is understandable that many of those responsible for delivering Early Warning Systems (EWS) might not wish to take the risk of implementing solutions perceived as containing unknown elements, despite the computational advantages that ANNs offer. This thesis therefore builds on existing efforts to open the box and develop tools and techniques that visualise, analyse and use ANN weights and biases especially from the viewpoint of neural pathways from inputs to outputs of feedforward networks. In so doing, it aims to demonstrate novel approaches to self-improving predictive model construction for both regression and classification problems. This includes Neural Pathway Strength Feature Selection (NPSFS), which uses ensembles of ANNs trained on differing subsets of data and analysis of the learnt weights to infer degrees of relevance of the input features and so build simplified models with reduced input feature sets. Case studies are carried out for prediction of flooding at multiple nodes in urban drainage networks located in three urban catchments in the UK, which demonstrate rapid, accurate prediction of flooding both for regression and classification. Predictive skill is shown to reduce beyond the time of concentration of each sewer node, when actual rainfall is used as input to the models. Further case studies model and predict statutory bacteria count exceedances for bathing water quality compliance at 5 beaches in Southwest England. An illustrative case study using a forest fires dataset from the UCI machine learning repository is also included. Results from these model ensembles generally exhibit improved performance, when compared with single ANN models. Also ensembles with reduced input feature sets, using NPSFS, demonstrate as good or improved performance when compared with the full feature set models. Conclusions are drawn about a new set of tools and techniques, including NPSFS and visualisation techniques for inspection of ANN weights, the adoption of which it is hoped may lead to improved confidence in the use of ANN for live real-time EWS applications.

Conselho Nacional de Desenvolvimento Científico e Tecnológico / As simulações computacionais tem sido amplamente empregadas no estudo do escoamento darciano e não-darciano em meios porosos consolidados e não-consolidados. Neste trabalho, através de uma nova formulação para a equação de Forchheimer, foram identificadas duas novas propriedades denominados fator de comportamento do fluido, que atua paralelamente a permeabilidade, e permeabilidade equivalente global, resultado da relação anterior. Este comportamento foi estudado e validado através da implementação de um aparato experimental e um código computacional baseado no modelo de regressão-linear que, além disso, demonstrou que o escoamento, ainda que em regime não darciano, comporta-se linearmente como a equação de Darcy, ainda que o coeficiente angular desta diminuiu de acordo com a faixa do número de Reynolds atingida, sendo esta dependente do tipo de leito empregado. Ainda neste trabalho, foi implementado o método de otimização R2W para estimar os parâmetros da equação de Kozeny-Carman a partir de dados experimentais obtidos por Dias et al, a fim de simular o escoamento darciano em meios porosos. Por fim, foi alcançada excelente concordância entre os dados simulados pelo método R2W / equação de Kozeny-Carman e os dados reais. / Computer simulations have been widely used in the study of Darcys flow and non-Darcy porous media in consolidated and non-consolidated. In this work, through a new formulation for the Forchheimer equation, we have been identified two new called Fluid Factor Behavior properties, which acts parallel to permeability, and overall equivalent permeability result of the previous relationship. This behavior has been studied and validated through implementation of an experimental apparatus and a computer code based on the linear regression model, moreover, it demonstrated that flow, even in non darciano system behaves linearly as the Darcy, however, the slope of this decreased according to the range of Reynolds numbers reached, this being dependent on the type of bed used. Although this work was implemented R2W optimization method to estimate the parameters of Kozeny-Carman equation from experimental data provided in the literature in order to simulate the darciano flow in porous media. Finally, it achieved excellent agreement between the data simulated by R2W method / Kozeny-Carman equation and actual data.

Dans le contexte du réchauffement et du changement climatique, il est important d’étudier les distributions, les cycles saisonniers et les tendances des gaz à l’état de trace dans l’atmosphère tels que la vapeur d’eau. L'Observatoire de Physique du Globe de Clermont-Ferrand a en charge plusieurs dispositifs d’observation dont le site instrumenté Cézeaux, Opme et Puy de Dôme (CO-PDD) situés dans le centre de la France (45◦ N, 3◦ E). Le site des Cézeaux dispose d’un LIDAR Rayleigh – Mie - Raman fournissant en routine des profils verticaux de vapeur d’eau et de paramètres optiques caractérisant les cirrus. Le site du puy de Dôme est équipé d’un spectroscope à cavité optique (CRDS-Picarro). Des mesures de colonnes totales de vapeur d’eau sont disponibles sur tous ces sites par GPS. Le site d’Opme est équipé d’un pluviomètre. Enfin Météo-France effectue le travail de contrôle qualité des données météorologiques sur les stations de mesure en France et ces données ont été utilisées dans cette thèse. La validation des données sur le site du puy de Dôme a été la première la première exploitation dans cette thèse. Des comparaisons des données sur le puy de Dôme ont montré un très bon accord entre les données de vapeur d’eau extraites de la station météorologique du puy de Dôme, de Météo France et les donnes CRDS du puy de Dôme, avec une corrélation de 0.94 et 0.98 respectivement. Les profils verticaux obtenus par LIDAR ont permis de sélectionner une anomalie de vapeur d’eau et d’identifier une intrusion stratosphère-troposphère en analysant les processus dynamique associés à cette anomalie. Les données OLR ont montré que cette intrusion est accompagnée de convection profonde et enfin LACYTRAJ confirme l'origine stratosphérique d’une partie de la masse d'air présente au-dessus de Clermont-Ferrand au cours de l’anomalie. Les longues séries d’observations (ex : Puy de Dôme 1995-2015) et des ré-analyse ECMWF ERA-Interim (1979-2017) et la diversité des sources de données (ex : satellites AIRS et COSMIC), nous permettent de documenter, analyser et comparer la variabilité, les cycles et la tendance de la vapeur d'eau à la surface et dans la troposphère, à différentes échelles de temps et déterminer les processus géophysiques responsables des distributions de vapeur d'eau sur le site CO-PDD. Le cycle annuel de la vapeur d'eau est clairement établi pour les deux sites de différentes altitudes et pour tous les types de mesure. Les sites de Cézeaux et du puy de Dôme ne présentent presque aucun cycle diurne, suggérant que la variabilité de la vapeur d'eau à la surface sur ces deux sites est plus influencée par les systèmes météorologique sporadiques que par les variations diurnes régulières. Les données LIDAR montrent une plus grande variabilité mensuelle de la distribution verticale que les produits satellites COSMIC et AIRS. La colonne totale de vapeur d'eau GPS sur le site des Cézeaux présente une tendance positive (0,42 ± 0,45 g/kg*décade entre 2006-2017). L'analyse par régressions multi-linéaires montre que les forçages continentaux (East Atlantic, East Atlantic-West Russia) ont une plus grande influence que le forçage océanique (Nord Atlantic Oscillation) sur les variations de vapeur d'eau. / In the context of global warming and climate change, it is important to study the distributions, seasonal cycles and trends of trace gases in the atmosphere such as water vapor. of the Observatoire de Physique du Globe de Clermont-Ferrand is in charge of several observation devices including the instrumented site Cézeaux, Opme and Puy de Dôme (CO-PDD) located near the center of France (45◦ N, 3◦ E). The site of Cézeaux is instrumented by a Rayleigh - Mie–LIDAR Raman providing routine vertical profiles of water vapor mixing ratio and optical parameters characterizing cirrus clouds. The puy de Dôme site is equipped with an optical cavity spectroscope (CRDS-Picarro). Measurements of total water vapor columns are available on all these sites by GPS. The Opme site is equipped with rain gauges. Finally, Météo-France performs the quality control work and of data on meteorological stations in France and these data were used in this thesis. The validation of the puy de Dôme data was the first the first task in this thesis. Comparisons between the puy de Dôme data sets showed a very good agreement between the water vapor datafrom the OPGC meteorological station of Puy de Dôme, Météo France and CRDS data with a correlation of 0.94 and 0.98 respectively. The vertical profiles deduced from the LIDAR allowed to identify a water vapor anomaly and a stratosphere-troposphere intrusion associated with this anomaly. OLR data showed that this intrusion could be linked with deep convection and LACYTRAJ confirms the stratospheric origin of a part of the air mass present above Clermont-Ferrand. Long series of observations (eg Puy de Dôme 1995-2015) and ECMWF ERA-Interim re-analysis (1979-2017) and the diversity of data sources (eg AIRS and COSMIC satellites), allowed us to document, analyze and compare the variability, cycles and trend of surface and tropospheric water vapor at different time scales and determine the geophysical processes responsible for water vapor distributions at the site of CO-PDD. The annual cycle of water vapor is clearly established for the two sites of different altitudes and for all types of measurement. Cézeaux and puy de Dôme present almost no diurnal cycle, suggesting that the variability of surface water vapor at this site is more influenced by a sporadic meteorological system than by regular diurnal variations. The LIDAR dataset shows a greater monthly variability of the vertical distribution than the COSMIC and AIRS satellite products. The Cézeaux site presents a positive trend for the GPS water vapor total column (0.42 ± 0.45 g/kg*decade during 2006–2017) and a significant negative trend for the surface water vapor mixing ratio (−0.16 ± 0.09 mm/decade during 2002–2017). The multi-linear regression analysis shows that continental forcings (East Atlantic Pattern and East Atlantic-West Russia Pattern) have a larger influence than oceanic forcing (North Atlantic Oscillation) on the water vapor variations.

Amphibians globally are declining with roughly a third facing extinction due primarily to threats linked to human impacts. One way in which this is being combated is by captive breeding programs. Nordens Ark in Sweden, in collaboration with other organizations, is attempting to breed one of the most threatened frogs for future reintroduction to the wild. The mountain chicken (Leptodactylus fallax) has unique characteristics related to its reproduction which make it challenging to breed successful offspring. Using Nordens Ark’s data on environmental conditions within breeding enclosures, this study attempts to determine which extrinsic factors are having the most influence on reproduction, in order to increase the success of the captive breeding program. Multiple linear regression analysis and model simplification using Stepwise regression and Akaike information criterion (AIC) were performed in order to determine which explanatory variables had the most influence in relation to foam nests. A reduced model with significant values for explanatory variables was deemed the best model based on the dataset. Of the 14 environmental variables tested, minimum temperature after nest construction had the most influence on foam nest length. When compared to wild habitats, temperatures within enclosures may be too high or lack seasonal cyclical patterns. Barometric pressure and humidity also had influence on foam nests, but the relationships between these environmental conditions is difficult to parse. Food supplements were also a significant factor and suggest that lack of vitamins or nutrients can have a negative effect on reproductive success. Ensuring the optimal conditions can be difficult for captive breeding programs, especially as amphibians are under-represented and face biases in conservation and research.

From the Proceedings of the 1974 Meetings of the Arizona Section - American Water Resources Assn. and the Hydrology Section - Arizona Academy of Science - April 19-20, 1974, Flagstaff, Arizona / This study attempts to determine the water quality changes in stock tanks and what factors are instrumental in the changes, and to assess the effects of the changes. Algal growth was the most prominent change taking place in the tanks with time. Little change in the water chemistry was noted until just before the tanks dried up. As algae died, ions tied up by the algae were released to the water, causing an increase in concentration of many of the nutrients. In order to determine the impact of various factors on algal growth, the data from eight stock tanks were analyzed by stepwise linear regression. Although 20 variables were used in the complete analysis, six variables were associated with 56.3 percent of the variance: time (since first sampling), total n, potassium, pH, inflow (recharge to the tanks), and hco3 concentration. Time and total n explained 51.3 percent of the variance, and potassium increased the variance to 52.8 percent. The pH reversed the relative positions of time and total n, with total n becoming dominant. The last two factors, inflow and hco3 were negative (resulting in a decrease in algal population) and increased the coefficient of variance to 56.3 percent.

From the Proceedings of the 1977 Meetings of the Arizona Section - American Water Resources Assn. and the Hydrology Section - Arizona Academy of Science - April 15-16, 1977, Las Vegas, Nevada / Increased nutritional and economic demands for agricultural products have dictated the need for greater and more efficient use of western grass forage. Vegetation manipulation is the quickest and most economical means of increasing forage. However , the hydrologic effects must be taken into consideration before embarking on a large scale vegetation manipulated program. This study discusses the hydrologic and erosion changes measured from a 110-acre semiarid watershed which was converted from brush to grass by root plowing and seeding. Significant changes were observed in rainfall-runoff relationships as average summer runoff was considerably in excess of predictions. Sediment yield also varied, and both of these results were tied to the change in vegetative cover and post conversion rainfall conditions.