Cluster Analysis of Imputed Financial Data Using an Augmentation-Based Algorithm

The human brain can solve highly abstract reasoning problems using a neural network that is entirely physical. The underlying mechanisms are only partially understood, but an artificial network provides valuable insight. See Article p.471 Conventional computer algorithms can process extremely large and complex data structures such as the worldwide web or social networks, but they must be programmed manually by humans. Neural networks can learn from examples to recognize complex patterns, but they cannot easily parse and organize complex data structures. Now Alex Graves, Greg Wayne and colleagues have developed a hybrid learning machine, called a differentiable neural computer (DNC), that is composed of a neural network that can read from and write to an external memory structure analogous to the random-access memory in a conventional computer. The DNC can thus learn to plan routes on the London Underground, and to achieve goals in a block puzzle, merely by trial and error—without prior knowledge or ad hoc programming for such tasks.

Product Data Management (PDM) system as primary component of the PLM concept has the goal of supporting product development by appropriate methods. Interactive visualization methods support various user groups (e.g. designers, project leaders, managers) in the interpretation and analysis of large and complex data structures in PDM systems which lead to improved decision making. This paper presents a novel approach for the application of visualization methods in PDM systems and a generic architecture for efficient data interpretation.

The overwhelming need to improve the quality of complex data structures in healthcare is more important than ever. Although data quality has been the point of interest in many studies, none of them has focused on the development of quantitative and explainable methods for data imputation. In this work, we propose a "smart" imputation workflow to address missing data across complex data structures in the context of in silico clinical trials. AI algorithms were utilized to produce high-quality virtual patient profiles. A search algorithm was then developed to extract the best virtual patient profiles through the definition of a profile matching score (PMS). A case study was conducted, where the real dataset was randomly contaminated with multiple missing values (e.g., 10 to 50%). In total, 10000 virtual patient profiles with less than 0.02 Kullback-Leibler (KL) divergence were produced to estimate the PMS distribution. The best generator achieved the lowest average squared absolute difference (0.4) and average correlation difference (0.02) with the real dataset highlighting its increased effectiveness for data imputation across complex clinical data structures.

To evaluate the accuracy of signal intensity (SI) analysis on unenhanced fat-suppressed T1-weighted MR images in the diagnosis of pancreatic disease and to compare subjective interpretation with different quantitative measurements. The pancreas was evaluated in 159 patients (86 normal and 73 with pancreatic disease) with spoiled gradient echo (GRE) T1-weighted fat saturation MR images. The relative SI of the pancreas to liver and spleen was quantitatively measured using regions of interest (ROIs) and qualitatively assessed by two independent observers. The mean values between a normal and an abnormal pancreas with pancreas-liver ratios of 0.14 +/- 0.37 vs. -0.32 +/- 0.24, respectively, and pancreas-spleen ratios of 0.89 +/- 0.55 vs. 0.02 +/- 0.43, respectively, were significantly different (P < 0.001). The pancreas-liver SI ratio was significantly better than the pancreas-spleen ratio throughout the disease group (area under the receiver operating characteristic (ROC) curve +/- SD; 0.92 +/- 0.02 for pancreas-liver vs. 0.86 +/- 0.03 for pancreas-spleen, P < 0.01), and after excluding cases of acute pancreatitis (0.96 +/- 0.02 for pancreas-liver vs. 0.89 +/- 0.03 for pancreas-spleen, P < 0.01). There was no statistically significant difference between quantitative and qualitative analysis (area under the ROC curve +/- SD; 0.93 +/- 0.02 vs. 0.93 +/- 0.02 for the entire disease group; excluding acute pancreatitis 0.96 +/- 0.02 vs 0.97 +/- 0.02) for the diagnosis of pancreatic disease when using liver as internal standard. The interobserver concordance was very good (kappa > 0.71). The sensitivity of visual liver comparison was 80% in the entire disease group and 91% after the cases of acute pancreatitis were excluded, while specificity was 93%. The pancreas-liver ratio is the best quantitative means of distinguishing normal from abnormal pancreas. Visual observation by experienced observers (qualitative measurement) was just as accurate as quantitative measurement. Detection of pancreatic pathology can be made with high accuracy by visually comparing the SI of the pancreas with that of the normal liver.

In financial risk-control scenarios, complex data structures and high costs of misclassification impose greater demands on classification models. To enhance model discrimination under noise interference and class imbalance, an imbalanced financial data classification model is developed by integrating adaptive denoising technology and evolutionary ensemble learning. At the data level, the model performs noise-aware weighting and minority oversampling to achieve sample balance. At the model level, a multi-objective particle swarm optimization algorithm was employed to evolve and optimize the weights of multiple sub-classifiers, forming a self-adaptive ensemble structure. Experiments conducted on three public datasets, Lending Club, Credit Card Fraud Detection, and Give Me Some Credit, showed that the model achieved F1-scores of 92.1%, 90.8%, and 88.4%, respectively. The model maintained high classification accuracy and stability under different data distributions and operating conditions. Further simulation tests demonstrate consistent convergence and strong robustness, confirming the applicability and generalization capability of the model in complex financial data environments.

In this research, a conceptual review was made for the application of data mining to the financial sector. In other words, it is summarized in empirical studies using financial data and data mining method. As a result of the general scanning, it was determined that data mining is used in many areas and it is widely used in empirical studies in the field of finance. When the financial studies using data mining methods are examined, it has been determined that a significant part of the studies in question are carried out in the field of risk management and financial failure. In the study, it was also revealed that artificial neural networks are the most used data mining methods in the field of finance. As a result of the general survey, it was determined that there is a need for research on credit scoring, abnormal return estimation, manipulation detection and portfolio management with data mining methods.

The rapidly growing field of proteomics and related applied sectors in the life sciences demands convenient methodologies for detecting and measuring the levels of specific proteins as well as for screening and analyzing for interacting protein systems. Materials utilized for such protein detection and measurement platforms should meet particular specifications which include ease-of-mass manufacture, biological stability, chemical functionality, cost effectiveness, and portability. Polymers can satisfy many of these requirements and are often considered as choice materials in various biological detection platforms. Therefore, tremendous research efforts have been made for developing new polymers both in macroscopic and nanoscopic length scales as well as applying existing polymeric materials for protein measurements. In this review article, both conventional and alternative techniques for protein detection are overviewed while focusing on the use of various polymeric materials in different protein sensing technologies. Among many available detection mechanisms, most common approaches such as optical, electrochemical, electrical, mass-sensitive, and magnetic methods are comprehensively discussed in this article. Desired properties of polymers exploited for each type of protein detection approach are summarized. Current challenges associated with the application of polymeric materials are examined in each protein detection category. Difficulties facing both quantitative and qualitative protein measurements are also identified. The latest efforts on the development and evaluation of nanoscale polymeric systems for improved protein detection are also discussed from the standpoint of quantitative and qualitative measurements. Finally, future research directions towards further advancements in the field are considered.

To price and risk-manage over-the-counter derivatives, financial institutions have to estimate counterparty default risks based on liquidly quoted credit default swap (CDS) rates. For the vast majority of counterparties, liquid CDS quotes are not available and proxy CDS rates need to be constructed. Existing methods ignore counterparty-specific default risks and can lead to arbitrage. The authors propose to construct CDS proxy rates by machine learning techniques to associate liquidly quoted CDS proxy names to illiquid ones on the basis of observable financial feature variables. A benchmarking exercise shows that the proposed method leads to significantly smaller estimation errors than do existing methods. The authors tested 126 classifiers coming from the eight most popular algorithms, rank-ordered performance, and investigated performance variations among and within the classifiers. In Part I, the authors present the methodology, review the different machine learning techniques, and report on cross-classifier performance. In Part II, they focus on parametrization and intra-classifier performance, investigate correlation effects, and perform a benchmarking exercise. This is a first systematic study of CDS proxy construction by machine learning techniques and a first classifier comparison study entirely based on financial market data. The techniques should be of interest for financial institutions seeking proxies for CDS rates or other financial variables.

The paper presents a typology of small-scale forest owners in Baden-Wurttemberg based on multiple criteria and derived using multivariate cluster analysis of long-term accountancy network data. Four distinct types of landholders are identified based on a combination of structural, financial and biophysical data. These groups fit well with the present knowledge on small-scale forest owners in Baden-Wurttemberg. In addition, the members of each group display clear differences in attitudes towards forestry, giving further support for the validity of the grouping. A comparison is made of this typology and typologies derived using the single criteria of forestry region, forest size class, cutting volume, proportion of coniferous trees, forest income and main source of income. This comparison demonstrates the advantages of using cluster analysis to identify types of small-scale forest owners in south-west Germany. No matter whether structural, natural or financial data are assessed, the multiple criteria typology produced by cluster analysis provides the highest percentage of statistically significant distinctions between the medians of the groups identified. A typology based on the single criterion of ‘regions’ provides the second best fit and has the advantage of being simple to develop. A more detailed comparison between these two typologies, whereby the differences between groups were analysed based on the links between individual groups in respect to individual variables, also indicated that the multiple criteria typology was superior.

PurposeKnowing financial and economic information beforehand benefits in planning and developing policies for every country especially for a developing country like Thailand and for other Asian countries. Unfortunately, missing data or non-response plays an essential role in many areas of studies including finance and economics. Eradication of missing data in a proper way before further analysis can gain remarkable outcomes and can be effective for planning policies. This review on the generalized regression estimators for population total can be applied to financial, economic and other data when missing data are present.Design/methodology/approachThe generalized regression estimators for estimating population total, including the variance estimators under unequal probability sampling without replacement with missing data are explored under the reverse framework. Applications to financial and economic data in Thailand are also reviewed.FindingsThe review of literatures related to the proposed estimator shows the best performance, giving smaller variances in all scenarios.Originality/valueThe generalized regression estimators can assist in estimating financial and economic data that contain missing values with different missing mechanisms and can be used in other applications which help gain more superior estimators.

Data and network embedding techniques are essential for representing complex data structures in a lower-dimensional space, aiding in tasks like data inference and network reconstruction by assigning nodes to concise representations while preserving the network's structure. The integration of Particle Swarm Optimization (PSO) with matrix factorization methods optimizes mapping functions and parameters during the embedding process, enhancing representation learning efficiency. Combining PSO with techniques like Deep Walk highlights its adaptability as a robust optimization tool for extracting meaningful representations from intricate data and network architectures. This collaboration significantly advances network inference and reconstruction methodologies by streamlining the representation of complex data structures. Leveraging PSO's optimization capabilities enables researchers to extract high-quality information from data networks, improving the accuracy of data inference outcomes. The amalgamation of PSO with data and network embedding methodologies not only enhances the quality of extracted information but also drives innovations in network analysis and related fields. This integration streamlines representation learning and advances network analysis methodologies, enabling more precise data inference and reconstruction. The adaptability and efficiency of PSO in extracting meaningful representations from complex data structures underscore its significance in advancing network inference and reconstruction techniques, contributing to the evolution of network analysis methodologies.

This paper presents the analysis of element deletion in different data structures, like linked lists, binary search trees, stacks, and queues. For random data sequence, we build the linked list and binary search tree and we compare the element deletion procedure. We find the much more stable performance of the binary search tree than that of the linked list. Additionally, we analyze the stack as a complex data structure and its basic operations. We present the element deletion with the basic operations for array stack implementation and we note the limitations. Also, we present the element deletion problem for a queue as a complex data structure, implemented with an array. We present the element deletion with the basic queue operations and we point out the limitations.

The digital platform can manage the documents effectively, because it can help us to organize, query, and update the documents. Also, we can query the documents on different mobile terminal devices with the help of the network. However, some documents have complex data structure, such as the Bible. In the Bible, each chapter has its own characteristic. The topic of each chapter in the Bible is introduced in its first sentence or its first paragraph. Therefore, it is necessary to use modern digital platform to organize complex data structure in the Bible. Then, this paper designs and develops a new system. This system is based on the Laravel framework with MVC design pattern. The data components of this system are divided into three parts: Model Layer, View Layer, and Controller Layer. The feature includes the architecture of interactive software system and the technology of Laravel routing engine. The result of experiment shows that the system has an excellent processing ability of complex data structure. By using the Laravel framework, both the time of development and flow of design can be reduced efficiently.

A Data-Parallel FP Compiler

Discovering clusters remains challenging when dealing with complex data structures, including those with varying densities, arbitrary shapes, weak separability, or the presence of noise. In this article, we propose a novel clustering algorithm called representative point-based clustering with neighborhood information (RPC-NI), which highlights the significance of neighborhood information often overlooked by existing clustering methods. The proposed algorithm first introduces a new local centrality metric that integrates both neighborhood density and topological convergence to identify core representative points, effectively capturing the structural characteristics of the data. Subsequently, a density-adaptive distance is defined to evaluate dissimilarities between these core representative points, and such distance is used to construct a minimum spanning tree (MST) over these points. Finally, an MST-based clustering algorithm is employed to yield the desired clusters. Incorporating neighborhood information enables RPC-NI to comprehensively determine representative points, and having multiple representative points per cluster allows RPC-NI to adapt to clusters of arbitrary shapes, varying densities, and different sizes. Extensive experiments on widely used datasets demonstrate that RPC-NI outperforms baseline algorithms in terms of clustering accuracy and robustness. These results provide further evidence for the importance of incorporating neighborhood information discovering clusters with complex structures.