A Dependency Parsing Approach to Biomedical Text Mining

Biomedical research is currently facing a new type of challenge: an excess of information, both in terms of raw data from experiments and in the number of scientific publications describing their results. Mirroring the focus on data mining techniques to address the issues of structured data, there has recently been great interest in the development and application of text mining techniques to make more effective use of the knowledge contained in biomedical scientific publications, accessible only in the form of natural human language. This thesis describes research done in the broader scope of projects aiming to develop methods, tools and techniques for text mining tasks in general and for the biomedical domain in particular. The work described here involves more specifically the goal of extracting information from statements concerning relations of biomedical entities, such as protein-protein interactions. The approach taken is one using full parsing—syntactic analysis of the entire structure of sentences—and machine learning, aiming to develop reliable methods that can further be generalized to apply also to other domains. The five papers at the core of this thesis describe research on a number of distinct but related topics in text mining. In the first of these studies, we assessed the applicability of two popular general English parsers to biomedical text mining and, finding their performance limited, identified several specific challenges to accurate parsing of domain text. In a follow-up study focusing on parsing issues related to specialized domain terminology, we evaluated three lexical adaptation methods. We found that the accurate resolution of unknown words can considerably improve parsing performance and introduced a domain-adapted parser that reduced the error rate of theoriginal by 10% while also roughly halving parsing time. To establish the relative merits of parsers that differ in the applied formalisms and the representation given to their syntactic analyses, we have also developed evaluation methodology, considering different approaches to establishing comparable dependency-based evaluation results. We introduced a methodology for creating highly accurate conversions between different parse representations, demonstrating the feasibility of unification of idiverse syntactic schemes under a shared, application-oriented representation. In addition to allowing formalism-neutral evaluation, we argue that such unification can also increase the value of parsers for domain text mining. As a further step in this direction, we analysed the characteristics of publicly available biomedical corpora annotated for protein-protein interactions and created tools for converting them into a shared form, thus contributing also to the unification of text mining resources. The introduced unified corpora allowed us to perform a task-oriented comparative evaluation of biomedical text mining corpora. This evaluation established clear limits on the comparability of results for text mining methods evaluated on different resources, prompting further efforts toward standardization. To support this and other research, we have also designed and annotated BioInfer, the first domain corpus of its size combining annotation of syntax and biomedical entities with a detailed annotation of their relationships. The corpus represents a major design and development effort of the research group, with manual annotation that identifies over 6000 entities, 2500 relationships and 28,000 syntactic dependencies in 1100 sentences. In addition to combining these key annotations for a single set of sentences, BioInfer was also the first domain resource to introduce a representation of entity relations that is supported by ontologies and able to capture complex, structured relationships. Part I of this thesis presents a summary of this research in the broader context of a text mining system, and Part II contains reprints of the five included publications.

Electrical machine in a Hybrid Electrical Vehicle

Hybrid electric vehicles (HEV) have attracted very much attention during the latest years. Increasing environmental concern and an increase in fuel prices are key factors for the growing interest towards the HEV. In a hybrid electric vehicle the power train consists of both a mechanical power system and an electric power transmission system. The major subsystems in the mechanical power system are the internal combustion engine which powers the vehicle; electric power transmission including an energy storage, power electronic inverter, hybrid control system; the electric motor drive that runs either in the generating mode or in the motoring mode to process the power flow between the energy storage and the electrical machine. This research includes two advanced electric motors for a parallel hybrid: induction machine and permanent magnets synchronous machine. In the thesis an induction motor and a permanent magnet motor are compared as propulsion motors. Electric energy storages are also studied.

Salient Pole Synchronous Machine Modelling in an Industrial Direct Torque Controlled Drive Application

Synchronous motors are used mainly in large drives, for example in ship propulsion systems and in steel factories' rolling mills because of their high efficiency, high overload capacity and good performance in the field weakening range. This, however, requires an extremely good torque control system. A fast torque response and a torque accuracy are basic requirements for such a drive. For large power, high dynamic performance drives the commonly known principle of field oriented vector control has been used solely hitherto, but nowadays it is not the only way to implement such a drive. A new control method Direct Torque Control (DTC) has also emerged. The performance of such a high quality torque control as DTC in dynamically demanding industrial applications is mainly based on the accurate estimate of the various flux linkages' space vectors. Nowadays industrial motor control systems are real time applications with restricted calculation capacity. At the same time the control system requires a simple, fast calculable and reasonably accurate motor model. In this work a method to handle these problems in a Direct Torque Controlled (DTC) salient pole synchronous motor drive is proposed. A motor model which combines the induction law based "voltage model" and motor inductance parameters based "current model" is presented. The voltage model operates as a main model and is calculated at a very fast sampling rate (for example 40 kHz). The stator flux linkage calculated via integration from the stator voltages is corrected using the stator flux linkage computed from the current model. The current model acts as a supervisor that prevents only the motor stator flux linkage from drifting erroneous during longer time intervals. At very low speeds the role of the current model is emphasised but, nevertheless, the voltage model always stays the main model. At higher speeds the function of the current model correction is to act as a stabiliser of the control system. The current model contains a set of inductance parameters which must be known. The validation of the current model in steady state is not self evident. It depends on the accuracy of the saturated value of the inductances. Parameter measurement of the motor model where the supply inverter is used as a measurement signal generator is presented. This so called identification run can be performed prior to delivery or during drive commissioning. A derivation method for the inductance models used for the representation of the saturation effects is proposed. The performance of the electrically excited synchronous motor supplied with the DTC inverter is proven with experimental results. It is shown that it is possible to obtain a good static accuracy of the DTC's torque controller for an electrically excited synchronous motor. The dynamic response is fast and a new operation point is achieved without oscillation. The operation is stable throughout the speed range. The modelling of the magnetising inductance saturation is essential and cross saturation has to be considered as well. The effect of cross saturation is very significant. A DTC inverter can be used as a measuring equipment and the parameters needed for the motor model can be defined by the inverter itself. The main advantage is that the parameters defined are measured in similar magnetic operation conditions and no disagreement between the parameters will exist. The inductance models generated are adequate to meet the requirements of dynamically demanding drives.

Groupware and Workflow Tools - A Case study of SAP Business Workflow Supporting System Integration in a Global Telecommunication Company

Workflow management systems aim at the controlled execution of complex application processes in distributed and heterogeneous environments. These systems will shape the structure of information systems in business and non-business environments. E business and system integration is a fertile soil for WF and groupware tools. This thesis aims to study WF and groupware tools in order to gather in house knowledge of WF to better utilize WF solutions in future, and to focus on SAP Business Workflow in order to find a global solution for Application Link Enabling support for system integration. Piloting this solution in Nokia collects the experience of SAP R/3 WF tool for other development projects in future. The literary part of this study will guide to the world of business process automation providing a general description of the history, use and potentials of WF & groupware software. The empirical part of this study begins with the background of the case study describing the IT environment initiating the case by introducing SAP R/3 in Nokia, the communication technique in use and WF tool. Case study is focused in one solution with SAP Business Workflow. This study provides a concept to monitor communication between ERP systems and to increase the quality of system integration. Case study describes a way create support model for ALE/EDI interfaces. Support model includes monitoring organization and the workflow processes to solve the most common IDoc related errors.

Performance Evaluation and Calibration of Machine Vision Measurement

This thesis presents the calibration and comparison of two systems, a machine vision system that uses 3 channel RGB images and a line scanning spectral system. Calibration. is the process of checking and adjusting the accuracy of a measuring instrument by comparing it with standards. For the RGB system self-calibrating methods for finding various parameters of the imaging device were developed. Color calibration was done and the colors produced by the system were compared to the known colors values of the target. Software drivers for the Sony Robot were also developed and a mechanical part to connect a camera to the robot was also designed. For the line scanning spectral system, methods for the calibrating the alignment of the system and the measurement of the dimensions of the line scanned by the system were developed. Color calibration of the spectral system is also presented.

Mechanical closing of the paperboard package and the effect of the shape of paperboard package to its manufacturability

The main objective of this thesis was to design a hinge and a closing mechanism for the plastic rim of a paperboard package. Of the hinge and closing mechanisms the 3D-models were designed using SolidWorks program and the functionality of the mechanisms was tested with rapid prototype models. When a mechanism that worked was found, the manufacturability of the mechanisms was tested in an injection molding machine with changeable inserts. Another objective of this thesis was to test the effect of the shape of paperboard package to its manufacturability. The effect of the packages shape was tested with plastic tools made for three different shaped trays. Suggestions for further research were made according to the results of the thesis.

Machine Learning and Clinical Text. Supporting Health Information Flow

Fluent health information flow is critical for clinical decision-making. However, a considerable part of this information is free-form text and inabilities to utilize it create risks to patient safety and cost-­effective hospital administration. Methods for automated processing of clinical text are emerging. The aim in this doctoral dissertation is to study machine learning and clinical text in order to support health information flow.First, by analyzing the content of authentic patient records, the aim is to specify clinical needs in order to guide the development of machine learning applications.The contributions are a model of the ideal information flow,a model of the problems and challenges in reality, and a road map for the technology development. Second, by developing applications for practical cases,the aim is to concretize ways to support health information flow. Altogether five machine learning applications for three practical cases are described: The first two applications are binary classification and regression related to the practical case of topic labeling and relevance ranking.The third and fourth application are supervised and unsupervised multi-class classification for the practical case of topic segmentation and labeling.These four applications are tested with Finnish intensive care patient records.The fifth application is multi-label classification for the practical task of diagnosis coding. It is tested with English radiology reports.The performance of all these applications is promising. Third, the aim is to study how the quality of machine learning applications can be reliably evaluated.The associations between performance evaluation measures and methods are addressed,and a new hold-out method is introduced.This method contributes not only to processing time but also to the evaluation diversity and quality. The main conclusion is that developing machine learning applications for text requires interdisciplinary, international collaboration. Practical cases are very different, and hence the development must begin from genuine user needs and domain expertise. The technological expertise must cover linguistics,machine learning, and information systems. Finally, the methods must be evaluated both statistically and through authentic user-feedback.

Concentrated Winding Multiphase Permanent Magnet Machine Design and Electromagnetic Properties – Case Axial Flux Machine

Concentrated winding permanent magnet machines and their electromagnetic properties are studied in this doctoral thesis. The thesis includes a number of main tasks related to the application of permanent magnets in concentrated winding open slot machines. Suitable analytical methods are required for the first design calculations of a new machine. Concentrated winding machines differ from conventional integral slot winding machines in such a way that adapted analytical calculation methods are needed. A simple analytical model for calculating the concentrated winding axial flux machines is provided. The next three main design tasks are discussed in more detail in the thesis. The magnetic length of the rotor surface magnet machines is studied, and it is shown that the traditional methods have to be modified also in this respect. An important topic in this study has been to evaluate and minimize the rotor permanent magnet Joule losses by using segmented magnets in the calculations and experiments. Determination of the magnetizing and leakage inductances for a concentrated winding machine and the torque production capability of concentrated winding machines with different pole pair numbers are studied, and the results are compared with the corresponding properties of integral slot winding machines. The thesis introduces a new practical permanent magnet motor type for industrial use. The special features of the machine are based on the option of using concentrated winding open slot constructions of permanent magnet synchronous machines in the normal speed ranges of industrial motors, for instance up to 3000 min-1, without excessive rotor losses. By applying the analytical equations and methods introduced in the thesis, a 37 kW 2400 min-1 12-slot 10-pole axial flux machine with rotor-surfacemounted magnets is designed. The performance of the designed motor is determined by experimental measurements and finite element calculations.

Computational and Experimental Investigation of the Grooved Roll in Paper Machine Environment

In the paper machine, it is not a desired feature for the boundary layer flows in the fabric and the roll surfaces to travel into the closing nips, creating overpressure. In this thesis, the aerodynamic behavior of the grooved roll and smooth rolls is compared in order to understand the nip flow phenomena, which is the main reason why vacuum and grooved roll constructions are designed. A common method to remove the boundary layer flow from the closing nip is to use the vacuum roll construction. The downside of the use of vacuum rolls is high operational costs due to pressure losses in the vacuum roll shell. The deep grooved roll has the same goal, to create a pressure difference over the paper web and keep the paper attached to the roll or fabric surface in the drying pocket of the paper machine. A literature review revealed that the aerodynamic functionality of the grooved roll is not very well known. In this thesis, the aerodynamic functionality of the grooved roll in interaction with a permeable or impermeable wall is studied by varying the groove properties. Computational fluid dynamics simulations are utilized as the research tool. The simulations have been performed with commercial fluid dynamics software, ANSYS Fluent. Simulation results made with 3- and 2-dimensional fluid dynamics models are compared to laboratory scale measurements. The measurements have been made with a grooved roll simulator designed for the research. The variables in the comparison are the paper or fabric wrap angle, surface velocities, groove geometry and wall permeability. Present-day computational and modeling resources limit grooved roll fluid dynamics simulations in the paper machine scale. Based on the analysis of the aerodynamic functionality of the grooved roll, a grooved roll simulation tool is proposed. The smooth roll simulations show that the closing nip pressure does not depend on the length of boundary layer development. The surface velocity increase affects the pressure distribution in the closing and opening nips. The 3D grooved roll model reveals the aerodynamic functionality of the grooved roll. With the optimal groove size it is possible to avoid closing nip overpressure and keep the web attached to the fabric surface in the area of the wrap angle. The groove flow friction and minor losses play a different role when the wrap angle is changed. The proposed 2D grooved roll simulation tool is able to replicate the grooved aerodynamic behavior with reasonable accuracy. A small wrap angle predicts the pressure distribution correctly with the chosen approach for calculating the groove friction losses. With a large wrap angle, the groove friction loss shows too large pressure gradients, and the way of calculating the air flow friction losses in the groove has to be reconsidered. The aerodynamic functionality of the grooved roll is based on minor and viscous losses in the closing and opening nips as well as in the grooves. The proposed 2D grooved roll model is a simplification in order to reduce computational and modeling efforts. The simulation tool makes it possible to simulate complex paper machine constructions in the paper machine scale. In order to use the grooved roll as a replacement for the vacuum roll, the grooved roll properties have to be considered on the basis of the web handling application.

Novel Robot Solutions for Carrying out Field Joint Welding and Machining in the Assembly of the Vacuum Vessel of ITER

It is necessary to use highly specialized robots in ITER (International Thermonuclear Experimental Reactor) both in the manufacturing and maintenance of the reactor due to a demanding environment. The sectors of the ITER vacuum vessel (VV) require more stringent tolerances than normally expected for the size of the structure involved. VV consists of nine sectors that are to be welded together. The vacuum vessel has a toroidal chamber structure. The task of the designed robot is to carry the welding apparatus along a path with a stringent tolerance during the assembly operation. In addition to the initial vacuum vessel assembly, after a limited running period, sectors need to be replaced for repair. Mechanisms with closed-loop kinematic chains are used in the design of robots in this work. One version is a purely parallel manipulator and another is a hybrid manipulator where the parallel and serial structures are combined. Traditional industrial robots that generally have the links actuated in series are inherently not very rigid and have poor dynamic performance in high speed and high dynamic loading conditions. Compared with open chain manipulators, parallel manipulators have high stiffness, high accuracy and a high force/torque capacity in a reduced workspace. Parallel manipulators have a mechanical architecture where all of the links are connected to the base and to the end-effector of the robot. The purpose of this thesis is to develop special parallel robots for the assembly, machining and repairing of the VV of the ITER. The process of the assembly and machining of the vacuum vessel needs a special robot. By studying the structure of the vacuum vessel, two novel parallel robots were designed and built; they have six and ten degrees of freedom driven by hydraulic cylinders and electrical servo motors. Kinematic models for the proposed robots were defined and two prototypes built. Experiments for machine cutting and laser welding with the 6-DOF robot were carried out. It was demonstrated that the parallel robots are capable of holding all necessary machining tools and welding end-effectors in all positions accurately and stably inside the vacuum vessel sector. The kinematic models appeared to be complex especially in the case of the 10-DOF robot because of its redundant structure. Multibody dynamics simulations were carried out, ensuring sufficient stiffness during the robot motion. The entire design and testing processes of the robots appeared to be complex tasks due to the high specialization of the manufacturing technology needed in the ITER reactor, while the results demonstrate the applicability of the proposed solutions quite well. The results offer not only devices but also a methodology for the assembly and repair of ITER by means of parallel robots.

Improving Cross Direction Basis Weight Profile in a Fourdrinier Paper Machine

The purpose of this study was to improve PM7’s basis weight CD profile in Stora Enso’s Berghuizer mill and to search mechanical defects which affect to the formation of the basis weight CD profile. In the theoretical part PM7’s structure was presented and the formation of the basis weight and caliper CD profiles was examined as well as disturbances which are affecting to the formation. The function of the control system was scrutinised for the side of CD profiles as well as the formation of the measured CD profiles. Tuning of the control system was examined through the response model and filtering. Specification of the response model and filtering was explained and how to determine 2sigma statistical number. In the end of the theoretical part ATPA hardware and a new profile browser were introduced. In the experimental part focus was in the beginning to search and remove mechanical defects which are affecting to CD profiles. The next step was to verify the reliability of the online measurements, to study the stability of the basis weight CD profile and to find out so called fingerprint, a basis weight CD profile which is unique for each paper machine. New response model and filtering value for basis weight CD profile was determined by bump tests. After a follow up period the affect of the new response model and filtering was analysed.

Measurement and Improvement of Wet Paper Web Strength

Various strength properties of paper are measured to tell how well it resists breaks in a paper machine or in printing presses. The most often measured properties are dry tensile strength and dry tear strength. However, in many situations where paper breaks, it is not dry. For example, in web breaks after the wet pressing the dry matter content can be around 45%. Thus, wet-web strength is often a more critical paper property than dry strength. Both wet and dry strength properties of the samples were measured with a L&W tensile tester. Originally this device was not designed for the measurement of the wet web tensile strength, thus a new procedure to handle the wet samples was developed. The method was tested with Pine Kraft (never dried). The effect of different strength additives on the wet-web and dry paper tensile strength was studied. The polymers used in this experiment were aqueous solution of a cationic polyamidoamine-epichlorohydrin resin (PAE), cationic hydrophilised polyisocyanate and cationic polyvinylamine (PVAm). From all three used chemicals only Cationic PAE considerably increased the wet web strength. However it was noticed that at constant solids content all chemicals decreased the wet web tensile strength. So, since all chemicals enhanced solid content it can be concluded that they work as drainage aids, not as wet web strength additives. From all chemicals only PVAm increased the dry strength and two other chemicals even decreased the strength. As chemicals were used in strong diluted forms and were injected into the pulp slurry, not on the surface of the papersheets, changes in samples densities did not happen. Also it has to be noted that all these chemicals are mainly used to improve the wet strength after the drying of the web.

Study of ligand-based virtual screening tools in computer-aided drug design

Virtual screening is a central technique in drug discovery today. Millions of molecules can be tested in silico with the aim to only select the most promising and test them experimentally. The topic of this thesis is ligand-based virtual screening tools which take existing active molecules as starting point for finding new drug candidates. One goal of this thesis was to build a model that gives the probability that two molecules are biologically similar as function of one or more chemical similarity scores. Another important goal was to evaluate how well different ligand-based virtual screening tools are able to distinguish active molecules from inactives. One more criterion set for the virtual screening tools was their applicability in scaffold-hopping, i.e. finding new active chemotypes. In the first part of the work, a link was defined between the abstract chemical similarity score given by a screening tool and the probability that the two molecules are biologically similar. These results help to decide objectively which virtual screening hits to test experimentally. The work also resulted in a new type of data fusion method when using two or more tools. In the second part, five ligand-based virtual screening tools were evaluated and their performance was found to be generally poor. Three reasons for this were proposed: false negatives in the benchmark sets, active molecules that do not share the binding mode, and activity cliffs. In the third part of the study, a novel visualization and quantification method is presented for evaluation of the scaffold-hopping ability of virtual screening tools.

A Survey and Classification of Software Testing Tools

Software testing is one of the essential parts in software engineering process. The objective of the study was to describe software testing tools and the corresponding use. The thesis contains examples of software testing tools usage. The study was conducted as a literature study, with focus on current software testing practices and quality assurance standards. In the paper a tool classifier was employed, and testing tools presented in study were classified according to it. We found that it is difficult to distinguish current available tools by certain testing activities as many of them contain functionality that exceeds scopes of a single testing type.

Simulating the performance of paper machine control

The topic of this thesis is the simulation of a combination of several control and data assimilation methods, meant to be used for controlling the quality of paper in a paper machine. Paper making is a very complex process and the information obtained from the web is sparse. A paper web scanner can only measure a zig zag path on the web. An assimilation method is needed to process estimates for Machine Direction (MD) and Cross Direction (CD) profiles of the web. Quality control is based on these measurements. There is an increasing need for intelligent methods to assist in data assimilation. The target of this thesis is to study how such intelligent assimilation methods are affecting paper web quality. This work is based on a paper web simulator, which has been developed in the TEKES funded MASI NoTes project. The simulator is a valuable tool in comparing different assimilation methods. The thesis contains the comparison of four different assimilation methods. These data assimilation methods are a first order Bayesian model estimator, an ARMA model based on a higher order Bayesian estimator, a Fourier transform based Kalman filter estimator and a simple block estimator. The last one can be considered to be close to current operational methods. From these methods Bayesian, ARMA and Kalman all seem to have advantages over the commercial one. The Kalman and ARMA estimators seems to be best in overall performance.