Indiana University-Purdue University Indianapolis (IUPUI) / Intrinsically disordered proteins (IDPs) are flexible proteins without defined 3D structures. Studies show that IDPs are abundant in nature and actively involved in numerous biological processes. Two crucial subjects in the study of IDPs lie in analyzing IDPs’ functions and identifying them. We thus carried out three projects to better understand IDPs. In the 1st project, we propose a method that separates IDPs into different function groups. We used the approach of CH-CDF plot, which is based the combined use of two predictors and subclassifies proteins into 4 groups: structured, mixed, disordered, and rare. Studies show different structural biases for each group. The mixed class has more order-promoting residues and more ordered regions than the disordered class. In addition, the disordered class is highly active in mitosis-related processes among others. Meanwhile, the mixed class is highly associated with signaling pathways, where having both ordered and disordered regions could possibly be important. The 2nd project is about identifying if an unknown protein is entirely disordered. One of the earliest predictors for this purpose, the charge-hydropathy plot (C-H plot), exploited the charge and hydropathy features of the protein. Not only is this algorithm simple yet powerful, its input parameters, charge and hydropathy, are informative and readily interpretable. We found that using different hydropathy scales significantly affects the prediction accuracy. Therefore, we sought to identify a new hydropathy scale that optimizes the prediction. This new scale achieves an accuracy of 91%, a significant improvement over the original 79%. In our 3rd project, we developed a per-residue C-H IDP predictor, in which three hydropathy scales are optimized individually. This is to account for the amino acid composition differences in three regions of a protein sequence (N, C terminus and internal). We then combined them into a single per-residue predictor that achieves an accuracy of 74% for per-residue predictions for proteins containing long IDP regions.

Eftersom samhället blir allt mer datadrivet revolutionerar användningen av AI och maskininlärning sättet företag fungerar och utvecklas på. Denna studie utforskar användningen av AI, Big Data och Natural Language Processing (NLP) för att förbättra affärsverksamhet och intelligens i företag. Huvudsyftet med denna avhandling är att undersöka om den nuvarande klassificeringsprocessen hos värdorganisationen kan upprätthållas med minskade driftskostnader, särskilt lägre moln-GPU-kostnader. Detta har potential att förbättra klassificeringsmetoden, förbättra produkten som företaget erbjuder sina kunder på grund av ökad klassificeringsnoggrannhet och stärka deras värdeerbjudande. Vidare utvärderas tre tillvägagångssätt mot varandra och implementationerna visar utvecklingen inom området. Modellerna som jämförs i denna studie inkluderar traditionella maskininlärningsmetoder som Support Vector Machine (SVM) och Logistisk Regression, tillsammans med state-of-the-art transformermodeller som BERT, både Pre-Trained och Fine-Tuned. Artikeln visar att det finns en avvägning mellan prestanda och kostnad vilket illustrerar problemet som många företag, som Valu8, står inför när de utvärderar vilket tillvägagångssätt de ska implementera. Denna avvägning diskuteras och analyseras sedan mer detaljerat för att utforska möjliga kompromisser från varje perspektiv i ett försök att hitta en balanserad lösning som kombinerar prestandaeffektivitet och kostnadseffektivitet. / As society is becoming more data-driven, Artificial Intelligence (AI) and Machine Learning are revolutionizing how companies operate and evolve. This study explores the use of AI, Big Data, and Natural Language Processing (NLP) in improving business operations and intelligence in enterprises. The primary objective of this thesis is to examine if the current classification process at the host company can be maintained with reduced operating costs, specifically lower cloud GPU costs. This can improve the classification method, enhance the product the company offers its customers due to increased classification accuracy, and strengthen its value proposition. Furthermore, three approaches are evaluated against each other, and the implementations showcase the evolution within the field. The models compared in this study include traditional machine learning methods such as Support Vector Machine (SVM) and Logistic Regression, alongside state-of-the-art transformer models like BERT, both Pre-Trained and Fine-Tuned. The paper shows a trade-off between performance and cost, showcasing the problem many companies like Valu8 stand before when evaluating which approach to implement. This trade-off is discussed and analyzed in further detail to explore possible compromises from each perspective to strike a balanced solution that combines performance efficiency and cost-effectiveness.

Indiana University-Purdue University Indianapolis (IUPUI) / This research performs a comparative study of techniques for combining spatial data and multivariate statistical methods for characterizing water quality conditions in a river basin. The study has been performed on the White River basin in central Indiana, and uses sixteen physical and chemical water quality parameters collected from 44 different monitoring sites, along with various spatial data related to land use – land cover, soil characteristics, terrain characteristics, eco-regions, etc. Various parameters related to the spatial data were analyzed using ArcHydro tools and were included in the multivariate analysis methods for the purpose of creating classification equations that relate spatial and spatio-temporal attributes of the watershed to water quality data at monitoring stations. The study compares the use of various statistical estimates (mean, geometric mean, trimmed mean, and median) of monitored water quality variables to represent annual and seasonal water quality conditions. The relationship between these estimates and the spatial data is then modeled via linear and non-linear multivariate methods. The linear statistical multivariate method uses a combination of principal component analysis, cluster analysis, and discriminant analysis, whereas the non-linear multivariate method uses a combination of Kohonen Self-Organizing Maps, Cluster Analysis, and Support Vector Machines. The final models were tested with recent and independent data collected from stations in the Eagle Creek watershed, within the White River basin. In 6 out of 20 models the Support Vector Machine more accurately classified the Eagle Creek stations, and in 2 out of 20 models the Linear Discriminant Analysis model achieved better results. Neither the linear or non-linear models had an apparent advantage for the remaining 12 models. This research provides an insight into the variability and uncertainty in the interpretation of the various statistical estimates and statistical models, when water quality monitoring data is combined with spatial data for characterizing general spatial and spatio-temporal trends.

Recently, Attention Deficit Hyperactive Disorder (ADHD) is getting a lot of attention mainly for two reasons. First, it is one of the most commonly found childhood behavioral disorders. Around 5-10% of the children all over the world are diagnosed with ADHD. Second, the root cause of the problem is still unknown and therefore no biological measure exists to diagnose ADHD. Instead, doctors need to diagnose it based on the clinical symptoms, such as inattention, impulsivity and hyperactivity, which are all subjective. Functional Magnetic Resonance Imaging (fMRI) data has become a popular tool to understand the functioning of the brain such as identifying the brain regions responsible for different cognitive tasks or analyzing the statistical differences of the brain functioning between the diseased and control subjects. ADHD is also being studied using the fMRI data. In this dissertation we aim to solve the problem of automatic diagnosis of the ADHD subjects using their resting state fMRI (rs-fMRI) data. As a core step of our approach, we model the functions of a brain as a connectivity network, which is expected to capture the information about how synchronous different brain regions are in terms of their functional activities. The network is constructed by representing different brain regions as the nodes where any two nodes of the network are connected by an edge if the correlation of the activity patterns of the two nodes is higher than some threshold. The brain regions, represented as the nodes of the network, can be selected at different granularities e.g. single voxels or cluster of functionally homogeneous voxels. The topological differences of the constructed networks of the ADHD and control group of subjects are then exploited in the classification approach. We have developed a simple method employing the Bag-of-Words (BoW) framework for the classification of the ADHD subjects. We represent each node in the network by a 4-D feature vector: node degree and 3-D location. The 4-D vectors of all the network nodes of the training data are then grouped in a number of clusters using K-means; where each such cluster is termed as a word. Finally, each subject is represented by a histogram (bag) of such words. The Support Vector Machine (SVM) classifier is used for the detection of the ADHD subjects using their histogram representation. The method is able to achieve 64% classification accuracy. The above simple approach has several shortcomings. First, there is a loss of spatial information while constructing the histogram because it only counts the occurrences of words ignoring the spatial positions. Second, features from the whole brain are used for classification, but some of the brain regions may not contain any useful information and may only increase the feature dimensions and noise of the system. Third, in our study we used only one network feature, the degree of a node which measures the connectivity of the node, while other complex network features may be useful for solving the proposed problem. In order to address the above shortcomings, we hypothesize that only a subset of the nodes of the network possesses important information for the classification of the ADHD subjects. To identify the important nodes of the network we have developed a novel algorithm. The algorithm generates different random subset of nodes each time extracting the features from a subset to compute the feature vector and perform classification. The subsets are then ranked based on the classification accuracy and the occurrences of each node in the top ranked subsets are measured. Our algorithm selects the highly occurring nodes for the final classification. Furthermore, along with the node degree, we employ three more node features: network cycles, the varying distance degree and the edge weight sum. We concatenate the features of the selected nodes in a fixed order to preserve the relative spatial information. Experimental validation suggests that the use of the features from the nodes selected using our algorithm indeed help to improve the classification accuracy. Also, our finding is in concordance with the existing literature as the brain regions identified by our algorithms are independently found by many other studies on the ADHD. We achieved a classification accuracy of 69.59% using this approach. However, since this method represents each voxel as a node of the network which makes the number of nodes of the network several thousands. As a result, the network construction step becomes computationally very expensive. Another limitation of the approach is that the network features, which are computed for each node of the network, captures only the local structures while ignore the global structure of the network. Next, in order to capture the global structure of the networks, we use the Multi-Dimensional Scaling (MDS) technique to project all the subjects from an unknown network-space to a low dimensional space based on their inter-network distance measures. For the purpose of computing distance between two networks, we represent each node by a set of attributes such as the node degree, the average power, the physical location, the neighbor node degrees, and the average powers of the neighbor nodes. The nodes of the two networks are then mapped in such a way that for all pair of nodes, the sum of the attribute distances, which is the inter-network distance, is minimized. To reduce the network computation cost, we enforce that the maximum relevant information is preserved with minimum redundancy. To achieve this, the nodes of the network are constructed with clusters of highly active voxels while the activity levels of the voxels are measured based on the average power of their corresponding fMRI time-series. Our method shows promise as we achieve impressive classification accuracies (73.55%) on the ADHD-200 data set. Our results also reveal that the detection rates are higher when classification is performed separately on the male and female groups of subjects. So far, we have only used the fMRI data for solving the ADHD diagnosis problem. Finally, we investigated the answers of the following questions. Do the structural brain images contain useful information related to the ADHD diagnosis problem? Can the classification accuracy of the automatic diagnosis system be improved combining the information of the structural and functional brain data? Towards that end, we developed a new method to combine the information of structural and functional brain images in a late fusion framework. For structural data we input the gray matter (GM) brain images to a Convolutional Neural Network (CNN). The output of the CNN is a feature vector per subject which is used to train the SVM classifier. For the functional data we compute the average power of each voxel based on its fMRI time series. The average power of the fMRI time series of a voxel measures the activity level of the voxel. We found significant differences in the voxel power distribution patterns of the ADHD and control groups of subjects. The Local binary pattern (LBP) texture feature is used on the voxel power map to capture these differences. We achieved 74.23% accuracy using GM features, 77.30% using LBP features and 79.14% using combined information. In summary this dissertation demonstrated that the structural and functional brain imaging data are useful for the automatic detection of the ADHD subjects as we achieve impressive classification accuracies on the ADHD-200 data set. Our study also helps to identify the brain regions which are useful for ADHD subject classification. These findings can help in understanding the pathophysiology of the problem. Finally, we expect that our approaches will contribute towards the development of a biological measure for the diagnosis of the ADHD subjects.

Statistical process control (SPC) applies the science of statistics to various process control in order to provide higher-quality products and better services. The K chart is one among the many important tools that SPC offers. Creation of the K chart is based on Support Vector Data Description (SVDD), a popular data classifier method inspired by Support Vector Machine (SVM). As any methods associated with SVM, SVDD benefits from a wide variety of choices of kernel, which determines the effectiveness of the whole model. Among the most popular choices is the Euclidean distance-based Gaussian kernel, which enables SVDD to obtain a flexible data description, thus enhances its overall predictive capability. This thesis explores an even more robust approach by incorporating the Mahalanobis distance-based kernel (hereinafter referred to as Mahalanobis kernel) to SVDD and compare it with SVDD using the traditional Gaussian kernel. Method's sensitivity is benchmarked by Average Run Lengths obtained from multiple Monte Carlo simulations. Data of such simulations are generated from multivariate normal, multivariate Student's (t), and multivariate gamma populations using R, a popular software environment for statistical computing. One case study is also discussed using a real data set received from Halberg Chronobiology Center. Compared to Gaussian kernel, Mahalanobis kernel makes SVDD and thus the K chart significantly more sensitive to shifts in mean vector, and also in covariance matrix.

Artificial Intelligence in Automotive Industry / Cost reduction is one of the main concerns in industry. Companies invest considerably for better performance in end-of-line fault diagnosis systems. A common strategy is to use data obtained from existing instrumentation. This research investigates the challenge of learning from historical data that have already been collected by companies. Machine learning is basically one of the most common and powerful techniques of artificial intelligence that can learn from data and identify fault features with no need for human interaction. In this research, labeled sound and vibration measurements are processed into fault signatures for vehicle starter motors and alternators. A fault detection and identification system has been developed to identify fault types for end-of-line testing of motors. However, labels are relatively difficult to obtain, expensive, time consuming and require experienced humans, while unlabeled samples needs less effort to collect. Thus, learning from unlabeled data together with the guidance of few labels would be a better solution. Furthermore, in this research, learning from unlabeled data with absolutely no human intervention is also implemented and discussed as well. / Thesis / Master of Applied Science (MASc)

Artificial intelligence techniques, such as artificial neural networks, random forests, or support vector machines, have been used to address a variety of problems in numerous industries. However, in many cases, models have to deal with issues such as imbalanced data or high multi-dimensionality. This thesis implements and compares the performance of support vector machines, random forests, and neural networks for a new bank account fraud detection, a use case defined by imbalanced data and high multi-dimensionality. The neural network achieved both the best AUC-ROC (0.889) and the best average precision (0.192). However, the results of the study indicate that the difference between the models’ performance is not statistically significant to reject the initial hypothesis that assumed equal model performances. / Artificiell intelligens, som artificiella neurala nätverk, random forests eller support vector machines, har använts för att lösa en mängd olika problem inom många branscher. I många fall måste dock modellerna hantera problem som obalanserade data eller hög flerdimensionalitet. Denna avhandling implementerar och jämför prestandan hos support vector machines, random forests och neurala nätverk för att upptäcka bedrägerier med nya bankkonton, ett användningsfall som definieras av obalanserade data och hög flerdimensionalitet. Det neurala nätverket uppnådde både den bästa AUC-ROC (0,889) och den bästa genomsnittliga precisionen (0,192). Resultaten av studien visar dock att skillnaden mellan modellernas prestanda inte är statistiskt signifikant för att förkasta den ursprungliga hypotesen som antog lika modellprestanda.

This report describes a quantitative analysis performed to compare two different methods on the task of pronoun resolution for Swedish. The first method, an implementation of Mitkov’s algorithm, is a heuristic-based method — meaning that the resolution is determined by a number of manually engineered rules regarding both syntactic and semantic information. The second method is data-driven — a Support Vector Machine (SVM) using dependency trees and word embeddings as features. Both methods are evaluated on an annotated corpus of Swedish news articles which was created as a part of this thesis. SVM-based methods significantly outperformed the implementation of Mitkov’s algorithm. The best performing SVM model relies on tree kernels applied to dependency trees. The model achieved an F1-score of 0.76 for the positive class and 0.9 for the negative class, where positives are pairs of pronoun and noun phrase that corefer, and negatives are pairs that do not corefer. / Rapporten beskriver en kvantitativ analys som genomförts för att jämföra två olika metoder för automatisk pronomenbestämning på svenska. Den första metoden, en implementation av Mitkovs algoritm, är en heuristisk metod vilket innebär att pronomenbestämningen görs med ett antal manuellt utformade regler som avser att fånga både syntaktisk och semantisk information. Den andra metoden är datadriven, en stödvektormaskin (SVM) som använder dependensträd och ordvektorer som särdrag. Båda metoderna utvärderades med hjälp av en annoterad datamängd bestående av svenska nyhetsartiklar som skapats som en del av denna avhandling. Den datadrivna metoden överträffade Mitkovs algoritm. Den SVM-modell som ger bäst resultat bygger på trädkärnor som tillämpas på dependensträd. Modellen uppnådde ett F1-värde på 0.76 för den positiva klassen och 0.9 för den negativa klassen, där de positiva datapunkterna utgörs av ett par av pronomen och nominalfras som korefererar, och de negativa datapunkterna utgörs av par som inte korefererar.

Sensor fault detection is an actively researched area and there are a plethora of studies on sensor fault detection in various applications such as nuclear power plants, wireless sensor networks, weather stations and nuclear fusion. However, there does not seem to be any study focusing on detecting sensor faults in the threaded fastening assembly application. Since the threaded fastening tools use torque and angle measurements to determine whether or not a screw or bolt has been fastened properly, faulty measurements from these sensors can have dire consequences. This study aims to investigate the use of machine learning to detect a subtle kind of sensor faults, common in this application, that are difficult to detect using canonical model-based approaches. Because of the subtle and infrequent nature of these faults, a two-stage system was designed. The first component of this system is given sensor data from a tightening and then tries to classify each data point in the sensor data as normal or faulty using a combination of low-pass filtering to generate residuals and a support vector machine to classify the residual points. The second component uses the output from the first one to determine if the complete tightening is normal or faulty. Despite the modest performance of the first component, with the best model having an F1-score of 0.421 for classifying data points, the design showed promising performance for classifying the tightening signals, with the best model having an F1-score of 0.976. These results indicate that there indeed exist patterns in these kinds of torque and angle multi-sensor signals that make machine learning a feasible approach to classify them and detect sensor faults. / Sensorfeldetektering är för nuvarande ett aktivt forskningsområde med mängder av studier om feldetektion i olika applikationer som till exempel kärnkraft, trådlösa sensornätverk, väderstationer och fusionskraft. Ett applikationsområde som inte verkar ha undersökts är det inom åtdragningsmontering. Eftersom verktygen inom åtdragningsmontering använder mätvärden på vridmoment och vinkel för att avgöra om en skruv eller bult har dragits åt tillräckligt kan felaktiga mätvärden från dessa sensorer få allvarliga konsekvenser. Målet med denna studie är att undersöka om det går att använda maskininlärning för att detektera en subtil sorts sensorfel som är vanlig inom åtdragningsmontering och har visat sig vara svåra att detektera med konventionella modell-baserade metoder. I och med att denna typ av sensorfel är både subtila och infrekventa designades ett system bestående av två komponenter. Den första får sensordata från en åtdragning och försöker klassificera varje datapunkt som antingen normal eller onormal genom att uttnyttja en kombination av lågpassfiltrering för att generera residualer och en stödvektormaskin för att klassificera dessa. Den andra komponenten använder resultatet från den första komponenten för att avgöra om hela åtdragningen ska klassificeras som normal eller onormal. Trots att den första komponenten hade ett ganska blygsamt resultat på att klassificera datapunkter så visade systemet som helhet mycket lovande resultat på att klassificera hela åtdragningar. Dessa resultat indikerar det finns mönster i denna typ av sensordata som gör maskininlärning till ett lämpligt verktyg för att klassificera datat och detektera sensorfel.

Data/information fusion is an integral component of many existing and emerging applications; e.g., remote sensing, smart cars, Internet of Things (IoT), and Big Data, to name a few. While fusion aims to achieve better results than what any one individual input can provide, often the challenge is to determine the underlying mathematics for aggregation suitable for an application. In this dissertation, I focus on the following three aspects of aggregation: (i) efficient data-driven learning and optimization, (ii) extensions and new aggregation methods, and (iii) feature and decision level fusion for machine learning with applications to signal and image processing. The Choquet integral (ChI), a powerful nonlinear aggregation operator, is a parametric way (with respect to the fuzzy measure (FM)) to generate a wealth of aggregation operators. The FM has 2N variables and N(2N − 1) constraints for N inputs. As a result, learning the ChI parameters from data quickly becomes impractical for most applications. Herein, I propose a scalable learning procedure (which is linear with respect to training sample size) for the ChI that identifies and optimizes only data-supported variables. As such, the computational complexity of the learning algorithm is proportional to the complexity of the solver used. This method also includes an imputation framework to obtain scalar values for data-unsupported (aka missing) variables and a compression algorithm (lossy or losselss) of the learned variables. I also propose a genetic algorithm (GA) to optimize the ChI for non-convex, multi-modal, and/or analytical objective functions. This algorithm introduces two operators that automatically preserve the constraints; therefore there is no need to explicitly enforce the constraints as is required by traditional GA algorithms. In addition, this algorithm provides an efficient representation of the search space with the minimal set of vertices. Furthermore, I study different strategies for extending the fuzzy integral for missing data and I propose a GOAL programming framework to aggregate inputs from heterogeneous sources for the ChI learning. Last, my work in remote sensing involves visual clustering based band group selection and Lp-norm multiple kernel learning based feature level fusion in hyperspectral image processing to enhance pixel level classification.