Data-driven clustering : new methods and applications

Abstract : This work is concerned with the development and application of novel unsupervised learning methods, having in mind two target applications: the analysis of forensic case data and the classification of remote sensing images. First, a method based on a symbolic optimization of the inter-sample distance measure is proposed to improve the flexibility of spectral clustering algorithms, and applied to the problem of forensic case data. This distance is optimized using a loss function related to the preservation of neighborhood structure between the input space and the space of principal components, and solutions are found using genetic programming. Results are compared to a variety of state-of--the-art clustering algorithms. Subsequently, a new large-scale clustering method based on a joint optimization of feature extraction and classification is proposed and applied to various databases, including two hyperspectral remote sensing images. The algorithm makes uses of a functional model (e.g., a neural network) for clustering which is trained by stochastic gradient descent. Results indicate that such a technique can easily scale to huge databases, can avoid the so-called out-of-sample problem, and can compete with or even outperform existing clustering algorithms on both artificial data and real remote sensing images. This is verified on small databases as well as very large problems. Résumé : Ce travail de recherche porte sur le développement et l'application de méthodes d'apprentissage dites non supervisées. Les applications visées par ces méthodes sont l'analyse de données forensiques et la classification d'images hyperspectrales en télédétection. Dans un premier temps, une méthodologie de classification non supervisée fondée sur l'optimisation symbolique d'une mesure de distance inter-échantillons est proposée. Cette mesure est obtenue en optimisant une fonction de coût reliée à la préservation de la structure de voisinage d'un point entre l'espace des variables initiales et l'espace des composantes principales. Cette méthode est appliquée à l'analyse de données forensiques et comparée à un éventail de méthodes déjà existantes. En second lieu, une méthode fondée sur une optimisation conjointe des tâches de sélection de variables et de classification est implémentée dans un réseau de neurones et appliquée à diverses bases de données, dont deux images hyperspectrales. Le réseau de neurones est entraîné à l'aide d'un algorithme de gradient stochastique, ce qui rend cette technique applicable à des images de très haute résolution. Les résultats de l'application de cette dernière montrent que l'utilisation d'une telle technique permet de classifier de très grandes bases de données sans difficulté et donne des résultats avantageusement comparables aux méthodes existantes.

Nitric oxide synthase (NOS) inhibition for one week improves renal sodium and water excretion in cirrhotic rats with ascites

Normalization of the increased vascular nitric oxide (NO) generation with low doses of NG-nitro-L-arginine methyl ester (L-NAME) corrects the hemodynamic abnormalities of cirrhotic rats with ascites. We have undertaken this study to investigate the effect of the normalization of vascular NO production, as estimated by aortic cyclic guanosine monophosphate (cGMP) concentration and endothelial nitric oxide synthase (eNOS) protein expression in the aorta and mesenteric artery, on sodium and water excretion. Rats with carbon tetrachloride-induced cirrhosis and ascites were investigated using balance studies. The cirrhotic rats were separated into two groups, one receiving 0.5 mg/kg per day of L-NAME (CIR-NAME) during 7 d, whereas the other group (CIR) was administrated the same volume of vehicle. Two other groups of rats were used as controls, one group treated with L-NAME and another group receiving the same volume of vehicle. Sodium and water excretion was measured on days 0 and 7. On day 8, blood samples were collected for electrolyte and hormone measurements, and aorta and mesenteric arteries were harvested for cGMP determination and nitric oxide synthase (NOS) immunoblotting. Aortic cGMP and eNOS protein expression in the aorta and mesenteric artery were increased in CIR as compared with CIR-NAME. Both cirrhotic groups had a similar decrease in sodium excretion on day 0 (0.7 versus 0.6 mmol per day, NS) and a positive sodium balance (+0.9 versus +1.2 mmol per day, NS). On day 7, CIR-NAME rats had an increase in sodium excretion as compared with the CIR rats (sodium excretion: 2.4 versus 0.7 mmol per day, P < 0.001) and a negative sodium balance (-0.5 versus +0.8 mmol per day, P < 0.001). The excretion of a water load was also increased after L-NAME administration (from 28+/-5% to 65+/-7, P < 0.05). Plasma renin activity, aldosterone and arginine vasopressin were also significantly decreased in the CIR-NAME, as compared with the CIR rats. The results thus indicate that normalization of aortic cGMP and eNOS protein expression in vascular tissue is associated with increased sodium and water excretion in cirrhotic rats with ascites.

Defining and searching for structural motifs using DeepView/Swiss-PdbViewer.

BACKGROUND: Today, recognition and classification of sequence motifs and protein folds is a mature field, thanks to the availability of numerous comprehensive and easy to use software packages and web-based services. Recognition of structural motifs, by comparison, is less well developed and much less frequently used, possibly due to a lack of easily accessible and easy to use software. RESULTS: In this paper, we describe an extension of DeepView/Swiss-PdbViewer through which structural motifs may be defined and searched for in large protein structure databases, and we show that common structural motifs involved in stabilizing protein folds are present in evolutionarily and structurally unrelated proteins, also in deeply buried locations which are not obviously related to protein function. CONCLUSIONS: The possibility to define custom motifs and search for their occurrence in other proteins permits the identification of recurrent arrangements of residues that could have structural implications. The possibility to do so without having to maintain a complex software/hardware installation on site brings this technology to experts and non-experts alike.

Advanced geostatistical and machine-learning models for spatial data analysis of radioactively contaminated regions

Radioactive soil-contamination mapping and risk assessment is a vital issue for decision makers. Traditional approaches for mapping the spatial concentration of radionuclides employ various regression-based models, which usually provide a single-value prediction realization accompanied (in some cases) by estimation error. Such approaches do not provide the capability for rigorous uncertainty quantification or probabilistic mapping. Machine learning is a recent and fast-developing approach based on learning patterns and information from data. Artificial neural networks for prediction mapping have been especially powerful in combination with spatial statistics. A data-driven approach provides the opportunity to integrate additional relevant information about spatial phenomena into a prediction model for more accurate spatial estimates and associated uncertainty. Machine-learning algorithms can also be used for a wider spectrum of problems than before: classification, probability density estimation, and so forth. Stochastic simulations are used to model spatial variability and uncertainty. Unlike regression models, they provide multiple realizations of a particular spatial pattern that allow uncertainty and risk quantification. This paper reviews the most recent methods of spatial data analysis, prediction, and risk mapping, based on machine learning and stochastic simulations in comparison with more traditional regression models. The radioactive fallout from the Chernobyl Nuclear Power Plant accident is used to illustrate the application of the models for prediction and classification problems. This fallout is a unique case study that provides the challenging task of analyzing huge amounts of data ('hard' direct measurements, as well as supplementary information and expert estimates) and solving particular decision-oriented problems.

Pathology and biology of peripheral T-cell lymphomas.

Peripheral T-cell lymphomas (PTCLs) represent a heterogeneous group of more than 20 neoplastic entities derived from mature T cells and natural killer (NK) cells involved in innate and adaptive immunity. With few exceptions these malignancies, which may present as disseminated, predominantly extranodal or cutaneous, or predominantly nodal diseases, are clinically aggressive and have a dismal prognosis. Their diagnosis and classification is hampered by several difficulties, including a significant morphological and immunophenotypic overlap across different entities, and the lack of characteristic genetic alterations for most of them. Although there is increasing evidence that the cell of origin is a major determinant for the delineation of several PTCL entities, however, the cellular derivation of most entities remains poorly characterized and/or may be heterogeneous. The complexity of the biology and pathophysiology of PTCLs has been only partly deciphered. In recent years, novel insights have been gained from genome-wide profiling analyses. In this review, we will summarize the current knowledge on the pathobiological features of peripheral NK/T-cell neoplasms, with a focus on selected disease entities manifesting as tissue infiltrates primarily in extranodal sites and lymph nodes.

Somatic hybridization between Citrus sinensis (L.) Osbeck and C. grandis (L.) Osbeck

This work had as objective to produce citrus somatic hybrids between sweet oranges and pummelos. After chemical fusion of sweet orange embryogenic protoplasts with pummelo mesophyll-derived protoplasts, plants were regenerated by somatic embryogenesis and acclimatized in a greenhouse. The hybrids of 'Hamlin' sweet orange + 'Indian Red' pummelo and 'Hamlin' sweet orange + 'Singapura' pummelo were confirmed by leaf morphology, chromosome counting and molecular analysis. These hybrids have potential to be used directly as rootstocks aiming blight, citrus tristeza virus, and Phytophthora-induced disease tolerance, as well as for rootstocks improvement programs.

Bemisia tabaci, Brevicoryne brassicae and Thrips tabaci abundance on Brassica oleracea var. acephala

Kale Brassica oleracea var. acephala is attacked by whitefly Bemisia tabaci, aphid Brevicoryne brassicae and Thrips tabaci. One of the main reasons for extensive insecticide application is the lack of information about factors that control insect population. The objectives of this study were to investigate the relationships between predators and parasitoids, organic compound leaves, levels of leaf nitrogen and potassium, total rainfall, relative humidity, sunlight and median temperature on the abundance of whitefly, aphid, and thrips in kale genotype "Talo Roxo". The beating tray method, direct counting and magnifying lens were used to estimate the number of these pests, predators and parasitoids. Median temperature, sunlight and relative humidity correlated to the amount of leaf nonacosane, which in turn was associated with aphids population increase. A tendency in the reduction of aphids and thrips populations with increase in total rainfall was observed. The whitefly can be a harmful pest in kale producing regions of higher temperature and smaller rainfall. In regions which present moderate temperature, where there is a high incidence of aphids, genotype with low leaf wax content should be chosen. Natural enemies, especially the parasitoid Adialytus spp., can control agents of the aphids population in kale.

On Automatic Diagnosis of Alzheimer's Disease based on Spontaneous Speech Analysis and Emotional Temperature

Alzheimer's disease is the most prevalent form of progressive degenerative dementia; it has a high socio-economic impact in Western countries. Therefore it is one of the most active research areas today. Alzheimer's is sometimes diagnosed by excluding other dementias, and definitive confirmation is only obtained through a post-mortem study of the brain tissue of the patient. The work presented here is part of a larger study that aims to identify novel technologies and biomarkers for early Alzheimer's disease detection, and it focuses on evaluating the suitability of a new approach for early diagnosis of Alzheimer’s disease by non-invasive methods. The purpose is to examine, in a pilot study, the potential of applying Machine Learning algorithms to speech features obtained from suspected Alzheimer sufferers in order help diagnose this disease and determine its degree of severity. Two human capabilities relevant in communication have been analyzed for feature selection: Spontaneous Speech and Emotional Response. The experimental results obtained were very satisfactory and promising for the early diagnosis and classification of Alzheimer’s disease patients.

Automated Traffic/Truck Weight Monitoring Equipment (Weigh-in-Motion), HR-282, 1988

This report documents work undertaken in the demonstration of a low-cost Automatic Weight and Classification System (AWACS). An AWACS procurement specification and details of the results of the project are also included. The intent of the project is to support and encourage transferring research knowledge to state and local agencies and manufacturers through field demonstrations. Presently available, Weigh-in-Motion and Classification Systems are typically too expensive to permit the wide deployment necessary to obtain representative vehicle data. Piezo electric technology has been used in the United Kingdom and Europe and is believed to be the basic element in a low-cost AWACS. Low-cost systems have been installed at two sites, one in Portland Cement Concrete (PCC) pavement in Iowa and the other in Asphaltic Cement Concrete (ACC) pavement in Minnesota to provide experience with both types of pavement. The systems provide axle weights, gross vehicle weight, axle spacing, vehicle classification, vehicle speed, vehicle count, and time of arrival. In addition, system self-calibration and a method to predict contact tire pressure is included in the system design. The study has shown that in the PCC pavement, the AWACS is capable of meeting the needs of state and federal highway agencies, producing accuracies comparable to many current commercial WIM devices. This is being achieved at a procurement cost of substantially less than currently available equipment. In the ACC pavement the accuracies were less than those observed in the PCC pavement which is concluded to result from a low pavement rigidity at this site. Further work is needed to assess the AWACS performance at a range of sites in ACC pavements.

Crushed Stone Granular Surfacing Materials, HR-2046 and MLR-90-07, 1990

This project included the following tasks: (1) Preparation of a questionnaire and survey of all 99 Iowa county engineers for input on current surfacing material practice; (2) County survey data analysis and selection of surfacing materials gradations to be used for test road construction; (3) Solicitation of county engineers and stone producers for project participation; (4) Field inspection and selection of the test road; (5) Construction of test road using varying material gradations from a single source; and (6) Field and laboratory testing and test road monitoring. The results of this research project indicate that crushed stone surfacing material graded on the fine side of Iowa Department of Transportation Class A surfacing specifications provides lower roughness and better rideability; better braking and handling characteristics; and less dust generation than the coarser gradations. It is believed that this material has sufficient fines available to act as a binder for the coarser material, which in turn promotes the formation of tight surface crust. This crust acts to provide a smooth riding surface, reduces dust generation, and improves vehicle braking and handling characteristics.

TraumaHawk: Improving Treatment for Crash Victims Using Photos and a Smartphone App

Photographic documentation of crashed vehicles at the scene can be used to improve triage of crash victims. A U.S. expert panel developed field triage rules to determine the likelihood of occupants sustaining serious injuries based on vehicle damage that would require transport to a trauma center (Sasser et al., 2011). The use of photographs for assessing vehicle damage and occupant compartment intrusion as it correlates to increased injury severity has been validated (Davidson et al., 2014). Providing trauma staff with crash scene photos remotely could assist them in predicting injuries. This would allow trauma care providers to assess the appropriate transport, as well as develop mental models of treatment options prior to patient arrival at the emergency department (ED). Crash-scene medical response has improved tremendously in the past 20-30 years. This is in part due to the increasing number of paramedics who now have advanced life support (ALS) training that allows independence in the field. However, while this advanced training provides a more streamlined field treatment protocol, it also means that paramedics focused on treating crash victims may not have time to communicate with trauma centers regarding crash injury mechanisms. As a result, trauma centers may not learn about severe trauma patients until just a few minutes before they arrive. The information transmitted by the TraumaHawk app allows interpretation of injury mechanisms from crash scene photos at the trauma center, providing clues about the type and severity of injury. With strategic crash scene photo documentation, trained trauma professionals can assess the severity and patterns of injury based on exterior crush and occupant intrusion. Intrusion increases the force experienced by vehicle occupants, which translates into a higher level of injury severity (Tencer et al., 2005; Assal et al., 2002; Mandell et al., 2010). First responders have the unique opportunity to assess the damaged vehicle at the crash scene, but often the mechanism of injury is limited or not even relayed to ED trauma staff. To integrate photographic and scene information, an app called TraumaHawk was created to capture images of crash vehicles and send them electronically to the trauma center. If efficiently implemented, it provides the potential advantage of increasing lead-time for preparation at the trauma center through the crash scene photos. Ideally, the result is better treatment outcomes for crash victims. The objective of this analysis was to examine if the extra lead-time granted by the TraumaHawk app could improve trauma team activation time over the current conventional communication method.

A Classification of Tools for Learning Logic

A table showing a comparison and classification of tools (intelligent tutoring systems) for e-learning of Logic at a college level.

Highway Maintenance Concept Vehicle Final Report: Phase Three, 2001

This report documents Phase III of a four-phase project. The goals of the project are to study the feasibility of using advanced technology from other industries to improve he efficiency and safety of winter highway maintenance vehicle operations, and to provide travelers with the level of service defined by policy during the winter season at the least cost to the taxpayers. The results of the first phase of the research were documented in the Concept Highway Maintenance Vehicle Final Report: Phase One dated April 1997, which describes the desirable functions of a concept maintenance vehicle and evaluates its feasibility. Phase I concluded by establishing the technologies that would be assembled and tested on the prototype vehicles in Phase II. The primary goals of phase II were to install the selected technologies on the prototype winter maintenance vehicles and to conduct proof of concept in advance of field evaluations planned for Phase III. This Phase III final report documents the work completed since the end of Phase II. During this time period, the Phase III work plan was completed and the redesigned friction meter was field tested. A vendor meeting was held to discuss future private sector participation and the new design for the Iowa vehicle. In addition, weather and roadway condition data were collected from the roadway weather information systems at selected sites in Iowa and Minnesota, for comparison to the vehicles' onboard temperature sensors. Furthermore, the team received new technology, such as the mobile Frensor unit, for bench testing and later installation.

Evaluating the Relationship between the Driver and Roadway to Address Rural Intersection Safety using the SHRP 2 Naturalistic Driving Study Data, RB05-013, 2016

Rural intersections account for 30% of crashes in rural areas and 6% of all fatal crashes, representing a significant but poorly understood safety problem. Transportation agencies have traditionally implemented countermeasures to address rural intersection crashes but frequently do not understand the dynamic interaction between the driver and roadway and the driver factors leading to these types of crashes. The Second Strategic Highway Research Program (SHRP 2) conducted a large-scale naturalistic driving study (NDS) using instrumented vehicles. The study has provided a significant amount of on-road driving data for a range of drivers. The present study utilizes the SHRP 2 NDS data as well as SHRP 2 Roadway Information Database (RID) data to observe driver behavior at rural intersections first hand using video, vehicle kinematics, and roadway data to determine how roadway, driver, environmental, and vehicle factors interact to affect driver safety at rural intersections. A model of driver braking behavior was developed using a dataset of vehicle activity traces for several rural stop-controlled intersections. The model was developed using the point at which a driver reacts to the upcoming intersection by initiating braking as its dependent variable, with the driver’s age, type and direction of turning movement, and countermeasure presence as independent variables. Countermeasures such as on-pavement signing and overhead flashing beacons were found to increase the braking point distance, a finding that provides insight into the countermeasures’ effect on safety at rural intersections. The results of this model can lead to better roadway design, more informed selection of traffic control and countermeasures, and targeted information that can inform policy decisions. Additionally, a model of gap acceptance was attempted but was ultimately not developed due to the small size of the dataset. However, a protocol for data reduction for a gap acceptance model was determined. This protocol can be utilized in future studies to develop a gap acceptance model that would provide additional insight into the roadway, vehicle, environmental, and driver factors that play a role in whether a driver accepts or rejects a gap.

Harmonization ofsignificant environmental aspects and their risk assessment in UPM paper and pulp mills

Työn tavoite onharmonisoida yhtenäiset rakenteet UPM:n paperi- ja sellutehtaiden merkittävilleympäristönäkökohdille sekä niiden ympäristöriskienhallintajärjestelmille. Näin saavutetaan yhteneväiset tavoitteet ja analysointikeinot yrityksen yksiköille. Harmonisointiprosessi on osa koko yrityksen ympäristöhallintajärjestelmän kehittämistä. Ja konsernin EMS -prosessi puolestaan konvergoi konsernin integroidun johtamisjärjestelmän kehitystä. Lisäksi työn tapaustutkimuksessa selvitettiin riskienhallintajärjestelmien integroitumispotentiaalia. Sen avulla saavutettaisiin paremmin suuren yrityksen synergia-etuja ja vuorovaikutteisuutta toimijoiden kesken, sekä parannettaisiin riskienhallintajärjestelmän mukautuvuutta ja käytettävyyttä. Työssä käsitellään kolmea esimerkkiä, joiden pohjalta tehdään esitys harmonisoiduille merkittäville ympäristönäkökohdille sekä riskienhallintajärjestelmien parametreille. Tutkimusongelmaa lähestytään haastattelujen, kirjallisuuden, yrityksen PWC:llä teettämän selvityksen sekä omien päätelmien avulla. Lisäksi työssä esitetään ympäristöhallintajärjestelmän tehokkuuden todentaminen ympäristösuorituskyvyn muuttujiin suhteutettuna. Pohjana jatkuvan kehityksen päämäärälle on organisaatio-oppiminen, niin yksittäisen työntekijän, tiimien kuin eri yksiköiden kesken. Se antaa sysäyksen aineettoman omaisuuden, kuten ympäristö-osaamisen, hyödyntämiseen parhaalla mahdollisella tavalla. Tärkeimpinä lopputuloksina työssä ovat ehdotukset harmonisoiduille merkittäville ympäristönäkökohdille sekä ympäristöriskienhallintajärjestelmän määritetyille komponenteille. Niitä ovat määritelmät ja skaalat riskien todennäköisyydelle, seurauksille sekä riskiluokille. Työn viimeisenä osana luodaan pohja tapaustutkimuksen avulla Rauman tehtaan jätevedenpuhdistamon kahden erilaisen riskienhallintajärjestelmän integroitumiselle.