Sviluppo ed applicazione di tecniche di Multiple Kernel Learning per dati di neuroimaging

Questo lavoro di tesi riguarda lo studio e l’implementazione di un algoritmo di multiple kernel learning (MKL) per la classificazione e la regressione di dati di neuroimaging ed, in particolare, di grafi di connettività funzionale. Gli algoritmi di MKL impiegano una somma pesata di vari kernel (ovvero misure di similarità) e permettono di selezionare le features utili alla discriminazione delle istanze durante l’addestramento del classificatore/regressore stesso. L’aspetto innovativo introdotto in questa tesi è stato lo studio di un nuovo kernel tra grafi di connettività funzionale, con la particolare caratteristica di conservare l’informazione relativa all’importanza di ogni singola region of interest (ROI) ed impiegando la norma lp come metodo per l’aggiornamento dei pesi, al fine di ottenere soluzioni sparsificate. L’algoritmo è stato validato utilizzando mappe di connettività sintetiche ed è stato applicato ad un dataset formato da 32 pazienti affetti da deterioramento cognitivo lieve e malattia dei piccoli vasi, di cui 16 sottoposti a riabilitazione cognitiva tra un’esame di risonanza ma- gnetica funzionale di baseline e uno di follow-up. Le mappe di con- nettività sono state ottenute con il toolbox CONN. Il classificatore è riuscito a discriminare i due gruppi di pazienti in una configurazione leave-one-out annidata con un’accuratezza dell’87.5%. Questo lavoro di tesi è stato svolto durante un periodo di ricerca presso la School of Computer Science and Electronic Engineering dell’University of Essex (Colchester, UK).

Characterisation of accessions and species of Macadamia to stem infection by Phytophthora cinnamomi

Phytophthora cinnamomi is a major pathogen in most macadamia plantations worldwide. Due to stem lesions, stem cankers and leaf defoliation it results in loss of productivity and tree death. In this study we examined accessions of the four Macadamia species and their hybrids, produced via rooted stem cuttings or germinated seeds, for susceptibility to stem canker and necrotic lesion caused by P. cinnamomi. Plants were wound-inoculated with agar containing P. cinnamomi. The symptoms produced in inoculated plants were used to characterize host susceptibility variation within and among the population. Lesion lengths and severity of stem canker were recorded. The four species and hybrids differed significantly in stem canker severity (P < 0.001) and lesion length (P = 0.04). M. integrifolia and M. tetraphylla hybrids were the most susceptible. M. integrifolia had the greatest stem canker severity and the most extensive lesions above and below the site of inoculation. Restricted lesion sizes were observed in M. ternifolia and M. jansenii. The effects of basal stem diameter and the method of propagation either from cuttings or seed were not significant. The genetic variation in the reactions of macadamia accessions to stem infection by P. cinnamomi is discussed.

Variation in susceptibility among macadamia genotypes and species to Phytophthora root decay caused by Phytophthora cinnamomi

Phytophthora cinnamomi is a major pathogen of cultivated macadamia (Macadamia integrifolia, Macadamia tetraphylla and their hybrids) worldwide. The susceptibility of the two non-edible Macadamia species (Macadamia ternifolia and Macadamia jansenii) to P. cinnamomi is not well-understood. Commercial macadamia trees are established on grafted seedling (seed propagation) or own-rooted cutting (vegetative propagation) rootstocks of hybrids of the cultivated species. There is little information to support the preferential use of rootstock propagated by either seedling or own-rooted cutting methods in macadamia. In this study we assessed roots of macadamia plants of the four species and their hybrids, derived from the two methods of propagation, for their susceptibility to P. cinnamomi infection. The roots of inoculated plant from which P. cinnamomi was recovered showed blackening symptoms. The non-cultivated species, M. ternifolia and M. jansenii and their hybrids were the most susceptible germplasm compared with M. tetraphylla and M. integrifolia. Of these two species, M. tetraphylla was less susceptible than M. integrifolia. Significant differences were observed among the accessions of their hybrids. A strong association (R2 > 0.75) was recorded between symptomatic roots and disease severity. Root density reduced with increasing disease severity rating in both own-rooted cuttings (R2 = 0.65) and germinated seedlings (R2 = 0.55). P. cinnamomi severity data were not significantly (P > 0.05) different between the two methods of plant propagation. The significance of this study to macadamia breeding and selection of disease resistant rootstocks is discussed.

Variation of cutting rooting ability in cultivated and wild species of Macadamia

In Australia, macadamia trees are commonly propagated by germinating rootstock seed and grafting when seedlings reach a suitable size. The production of grafted trees is a protracted and complex process, however, propagation of macadamia via cuttings represents a simpler and faster method of multiplication. Macadamias have traditionally proven difficult to propagate from cuttings, and while recent developments in the process have improved success rates, substantial variation in rooting ability between cultivars and species has been reported. The cultivar 'Beaumont' (Macadamia integrifolia × M. tetraphylla) is commonly propagated by cutting for use as a rootstock, and is relatively easy to strike while other cultivars are more difficult. There is speculation that Hawaiian cultivars are more difficult to strike from cuttings than Australian cultivars due to species and genetic composition. In this experiment, cuttings of 32 genotypes were evaluated for rooting ability. Each genotype's species profile was estimated using historical data, and used to determine species effects on survival (percentage) and rooting ability (rating 0-2). M. jansenii (100%), M. tetraphylla (84%) and M. integrifolia/tetraphylla hybrids (79%) had the highest success rates while M. integrifolia (54%) and M. ternifolia (43%) had the lowest survival. Rooting ability of M. jansenii (1.75) was significantly higher than M. ternifolia (0.49) but not significantly higher than M. tetraphylla × M. integrifolia with (1.09), M. tetraphylla (1.03) or M. integrifolia (0.88).

'Abnormal vertical growth' : A disorder threatening the viability of the Australian macadamia industry

'Abnormal vertical growth' (AVG) was recognised in Australia as a dysfunction of macadamia (Macadamia spp.) in the mid-1990s. Affected trees displayed unusually erect branching, and poor flowering and yield. Since 2002, the commercial significance of AVG, its cause, and strategies to alleviate its affects, has been studied. The cause is still unknown, and AVG remains a serious threat to orchard viability. AVG affects both commercial and urban macadamia. It occurs predominantly in the warmer-drier production regions of Queensland and New South Wales. An estimated 100,000 orchard trees are affected, equating to an annual loss of $ 10.5 M. In orchards, AVG occurs as aggregations of affected trees, affected tree number can increase by 4.5% per year, and yield reduction can exceed 30%. The more upright cultivars 'HAES 344' and '741' are highly susceptible, while the more spreading cultivars 'A4', 'A16' and 'A268' show tolerance. Incidence is higher (p<0.05) in soils of high permeability and good drainage. No soil chemical anomaly has been found. Fine root dry weight of AVG trees (0-15 cm depth) was found lower (p<0.05) than non-AVG. Next generation sequencing has led to the discovery of a new Bacillus sp. and a bipartite Geminivirus, which may have a role in the disease. Trunk cinctures will increase (p<0.05) yield of moderately affected trees. Further research is needed to clarify whether a pathogen is the cause, the role of soil moisture in AVG, and develop a varietal solution.

Assessing nut germination protocols for macadamia cultivar 'Beaumont'

A rapid rate and high percentage of macadamia nut germination, together with production of vigorous seedlings, are required by nurseries and breeding programs. Germination of nuts is typically protracted, however, and rarely reaches 100%. Many studies have been conducted into macadamia germination, but most have assessed percent germination only. This study investigated the effects of various treatments on percent germination, germination rate, and plant, shoot and root dry weights. The treatments tested were combinations of: (i) soaking or not soaking seeds in a dilute fungicide solution prior to planting; (ii) four different planting media; and (iii) leaving seed trays open or placing them inside clear plastic bags. For freshly harvested nuts, sowing in potting mix under clear plastic and without soaking produced the highest percent germination and germination rate, the largest shoots, and longest lateral roots.

Preliminary evaluation of macadamia rootstocks for yield and tree height

Rootstock has profound effects on traits such as yield and tree size in various horticultural industries, however relatively little is known about rootstock effects for macadamia. In this study, 12 cultivars were propagated as open-pollinated seedling and clonal rootstocks, and own-rooted cuttings. The same cultivars were also used as scions, and grafted to a subset of rootstocks, then planted at four trial locations. In this preliminary analysis, rootstock accounted for 19% of the variance in yield compared with 72% for scion, and 23% in height compared with 72% for scion. There was no interaction between rootstock and scion for yield, and only a small effect for height. The interaction between rootstock and propagation method (seedling, clonal, own roots) was not significant for height. A small effect was observed for yield, with the own roots treatment producing significantly lower yield than grafted trees for all rootstock cultivars except 'HAES 849'. 'H2' seedling rootstock produced a cumulative yield to age 10 years of 11.1 kg tree -1 compared to the highest yield of 13.6 kg tree -1 for 'Beaumont' clonal rootstocks. 'H2' seedling rootstock produced 4.8 m trees at age 11 years, compared to the smallest grafted tree which was 'HAES 849' seedling at 4.7 m.

Statistical forecasting of the Australian macadamia crop

Three types of forecasts of the total Australian production of macadamia nuts (t nut-in-shell) have been produced early each year since 2001. The first is a long-term forecast, based on the expected production from the tree census data held by the Australian Macadamia Society, suitably scaled up for missing data and assumed new plantings each year. These long-term forecasts range out to 10 years in the future, and form a basis for industry and market planning. Secondly, a statistical adjustment (termed the climate-adjusted forecast) is made annually for the coming crop. As the name suggests, climatic influences are the dominant factors in this adjustment process, however, other terms such as bienniality of bearing, prices and orchard aging are also incorporated. Thirdly, industry personnel are surveyed early each year, with their estimates integrated into a growers and pest-scouts forecast. Initially conducted on a 'whole-country' basis, these models are now constructed separately for the six main production regions of Australia, with these being combined for national totals. Ensembles or suites of step-forward regression models using biologically-relevant variables have been the major statistical method adopted, however, developing methodologies such as nearest-neighbour techniques, general additive models and random forests are continually being evaluated in parallel. The overall error rates average 14% for the climate forecasts, and 12% for the growers' forecasts. These compare with 7.8% for USDA almond forecasts (based on extensive early-crop sampling) and 6.8% for coconut forecasts in Sri Lanka. However, our somewhatdisappointing results were mainly due to a series of poor crops attributed to human reasons, which have now been factored into the models. Notably, the 2012 and 2013 forecasts averaged 7.8 and 4.9% errors, respectively. Future models should also show continuing improvement, as more data-years become available.

A model of macadamia with application to pruning in orchards

A self-organising model of macadamia, expressed using L-Systems, was used to explore aspects of canopy management. A small set of parameters control the basic architecture of the model, with a high degree of self-organisation occurring to determine the fate and growth of buds. Light was sensed at the leaf level and used to represent vigour and accumulated basipetally. Buds also sensed light so as to provide demand in the subsequent redistribution of the vigour. Empirical relationships were derived from a set of 24 completely digitised trees after conversion to multiscale tree graphs (MTG) and analysis with the OpenAlea software library. The ability to write MTG files was embedded within the model so that various tree statistics could be exported for each run of the model. To explore the parameter space a series of runs was completed using a high-throughput computing platform. When combined with MTG generation and analysis with OpenAlea it provided a convenient way in which thousands of simulations could be explored. We allowed the model trees to develop using self-organisation and simulated cultural practices such as hedging, topping, removal of the leader and limb removal within a small representation of an orchard. The model provides insight into the impact of these practices on potential for growth and the light distribution within the canopy and to the orchard floor by coupling the model with a path-tracing program to simulate the light environment. The lessons learnt from this will be applied to other evergreen, tropical fruit and nut trees.

Dry flower disease of Macadamia in Australia caused by Neopestalotiopsis macadamiae sp. nov. and Pestalotiopsis macadamiae sp. nov

Incidence of dry flower disease of macadamia (Macadamia integrifolia), expressed as blight of the flowers, necrosis and dieback of the rachis, is increasing in Australia. In the 2012/13 production season, incidence of dry flower disease resulted in 10% to 30% yield loss in the affected orchards. Etiology of the disease has not been established. This study was established to characterise the disease and identify the causal pathogen. A survey of the major macadamia producing regions in Australia revealed dry flower disease symptoms, regardless of cultivar or location at all stages of raceme development. Based on colony and conidial morphology, the majority (41%) of fungal isolates obtained from tissue samples were identified as Pestalotiopsis and Neopestalotiopsis spp. The phylogeny of the combined partial sequence of the internal transcribed spacer, beta-tubulin and translation elongation factor 1-alpha gene loci, segregated the isolates into two well supported clades, independent of location or part of the inflorescence affected. Further morphological examination supported the establishment of two new species, which are formally described as Neopestalotiopsis macadamiae sp. nov. and Pestalotiopsis macadamiae sp. nov. Using spore suspensions of isolates of both species, Koch?s postulates were fulfilled on three macadamia cultivars at all stages of raceme development. To our knowledge, this is the first report of species of Neopestalotiopsis and Pestalotiopsis as causal agents of inflorescence disease in macadamia.

Bahadur representation of linear kernel quantile estimator of VaR under mixing assumptions

Bahadur representation and its applications have attracted a large number of publications and presentations on a wide variety of problems. Mixing dependency is weak enough to describe the dependent structure of random variables, including observations in time series and longitudinal studies. This note proves the Bahadur representation of sample quantiles for strongly mixing random variables (including ½-mixing and Á-mixing) under very weak mixing coe±cients. As application, the asymptotic normality is derived. These results greatly improves those recently reported in literature.

Diseño e implementación del kernel de xDEVS, versión distribuida

Para entender nuestro proyecto, debemos comprender DEVS. Dentro de los formalismos más populares de representación de sistemas de eventos discretos se encuentra DES. En la década de los 70, el matemático Bernard Zeigler propuso un formalismo general para la representación de dichos sistemas. Este formalismo denominado DEVS (Discrete EVent System Specification) es el formalismo más general para el tratamiento de DES. DEVS permite representar todos aquellos sistemas cuyo comportamiento pueda describirse mediante una secuencia de eventos discretos. Estos eventos se caracterizan por un tiempo base en el que solo un número de eventos finitos puede ocurrir. DEVS Modelado y Simulación tiene múltiples implementaciones en varios lenguajes de programación como por ejemplo en Java, C# o C++. Pero surge la necesidad de implementar una plataforma distribuida estable para proporcionar la mecánica de interoperabilidad e integrar modelos DEVS diversificados. En este proyecto, se nos dará como código base el core de xDEVS en java, aplicado de forma secuencial y paralelizada. Nuestro trabajo será implementar el core de manera distribuida de tal forma que se pueda dividir un sistema DEVS en diversas máquinas. Para esto hemos utilizado sockets de java para hacer la transmisión de datos lo más eficiente posible. En un principio deberemos especificar el número de máquinas que se conectarán al servidor. Una vez estas se hayan conectado se les enviará el trabajo específico que deberán simular. Cabe destacar que hay dos formas de dividir un sistema DEVS las cuales están implementadas en nuestro proyecto. La primera es dividirlo en módulos atómicos los cuales son subsistemas indivisibles en un sistema DEVS. Y la segunda es dividir las funciones de todos los subsistemas en grupos y repartirlos entre las máquinas. En resumen el funcionamiento de nuestro sistema distribuido será comenzar ejecutando el trabajo asignado al primer cliente, una vez finalizado actualizará la información del servidor y este mandara la orden al siguiente y así sucesivamente.

SCALABLE APPROXIMATION OF KERNEL FUZZY C-MEANS

Virtually every sector of business and industry that uses computing, including financial analysis, search engines, and electronic commerce, incorporate Big Data analysis into their business model. Sophisticated clustering algorithms are popular for deducing the nature of data by assigning labels to unlabeled data. We address two main challenges in Big Data. First, by definition, the volume of Big Data is too large to be loaded into a computer’s memory (this volume changes based on the computer used or available, but there is always a data set that is too large for any computer). Second, in real-time applications, the velocity of new incoming data prevents historical data from being stored and future data from being accessed. Therefore, we propose our Streaming Kernel Fuzzy c-Means (stKFCM) algorithm, which reduces both computational complexity and space complexity significantly. The proposed stKFCM only requires O(n2) memory where n is the (predetermined) size of a data subset (or data chunk) at each time step, which makes this algorithm truly scalable (as n can be chosen based on the available memory). Furthermore, only 2n2 elements of the full N × N (where N >> n) kernel matrix need to be calculated at each time-step, thus reducing both the computation time in producing the kernel elements and also the complexity of the FCM algorithm. Empirical results show that stKFCM, even with relatively very small n, can provide clustering performance as accurately as kernel fuzzy c-means run on the entire data set while achieving a significant speedup.