Search

Since predictions of property values require an integration of a great deal of data, machine learning can play an important role. There are many factors influencing the sale price of private properties. In this paper, we focus on how different types of crimes affect the residential property sale price in an urban county in the USA - Pierce County, Washington. We are interested in what machine learning algorithms can be more effective in predicting the sales values. We worked on two sets of data. The first data source includes the physical attributes of a property, such as the square footage, quality, the year built and/or remodeled. The second data source contains the crime data for Pierce County from July 2018 to July 2019. We combined these two data sources into one full dataset. Then, we divided the full dataset into three clusters using the EM (Expectation Maximization) algorithm. The full and three clustered datasets were used to build two sets of data mining models -- one with crime data, and the other without them. We built our models using three algorithms -- decision trees, artificial neural networks, and random forests. In total, we created 32 models and evaluated them by calculating prediction errors of each model. We found that the random forest models produced the lowest values of errors. Using those models, we concluded that crime is a significant factor in predicting the sale price of residential properties in Pierce County, Washington. [ABSTRACT FROM AUTHOR]

Abstract: In this paper, we present new approaches to handling drift and shift in on-line data streams with the help of evolving fuzzy systems (EFS), which are characterized by the fact that their structure (rule base and parameters) is not fixed and not pre-determined, but is extracted from data streams on-line and in an incremental manner. When dealing with so-called drifts and s hifts in data streams, one needs to take into account (1) automatic detection of drifts and shifts and (2) automatic reaction to the drifts and shifts. This is important to avoid interruptions in the learning process and downtrends in predictive accuracy. To address the first problem, we propose an approach based on the concept fuzzy rule age. The second problem is addressed by including gradual forgetting of (1) antecedent parts and (2) consequent parameters. The latter can be achieved by including a forgetting factor in the recursive local learning process of the parameters, whose value is automatically extracted based on the intensity of the shift/drift. For addressing the former problem, we introduce two alternative methods: one is based on the evolving density-based clustering (eClustering) used to form the antecedents in the eTS approach; the other is based on the automatic adaptation of the learning rate of the evolving vector quantization (eVQ) method used to form the antecedent in the FLEXFIS approach. The paper concludes with an empirical evaluation of the impact of the proposed approaches in (on-line) real-world data sets in which drifts and shifts occur. [ABSTRACT FROM AUTHOR]

Abstract: Two novel approaches to the problem of brain signals (electroencephalogram (EEG)) classification are introduced in the paper. The first method is based on a modular probabilistic network architecture that employs multiple dependant hidden Markov models (DM-HMM-D) on the input features (channels). The second method, eClass, is based on an on-line evolvable fuzzy rule base of EEG signal prototypes that represent each class and take into consideration the spatial proximity between input signals. Both approaches use supervised learning but differ in their mode of operation. eClass is designed recursively, on-line, and has an evolvable structure, while DM-HMM-D is trained off-line, in a block-based mode, and has a fixed architecture. Both methods have been extensively tested on real EEG data that is recorded during several experimental sessions involving a single female subject who is exposed to mild pain induced by a laser beam. Experimental results illustrate the viability of the proposed approaches and their potential in solving similar classification problems. [Copyright &y& Elsevier]

This paper proposes a new extended zero-order Autonomous Learning Multiple-Model (ALMMo-0*) neuro-fuzzy approach in order to classify different heart disorders through sounds. ALMMo-0* is build upon the recently introduced ALMMo-0. In this paper ALMMo-0 is extended by adding a pre-processing structure which improves the performance of the proposed method. ALMMo-0* has as a learning engine composed of hierarchical a massively parallel set of 0-order fuzzy rules, which are able to self-adapt and provide transparent and human understandable IF... THEN representation. The heart sound recordings considered in the analysis were sourced from several contributors around the world. Data were collected from both clinical and nonclinical environment, and from healthy and pathological patients. Differently from mainstream machine learning approaches, ALMMo-0* is able to learn from unseen data. The main goal of the proposed method is to provide highly accurate models with high transparency, interpretability, and explainability for heart disorder diagnosis. Experiments demonstrated that the proposed neuro-fuzzy-based modeling is an efficient framework for these challenging classification tasks surpassing its state-of-the-art competitors in terms of classification accuracy. Additionally, ALMMo-0* produced transparent AnYa type fuzzy rules, which are human interpretable, and may help specialists to provide more accurate diagnosis. Medical doctors can easily identify abnormal heart sounds by comparing a patient's sample with the identified prototypes from abnormal samples by ALMMo-0*. • Classification of heart sounds using a zero-order autonomous learning multiple-model classifier. • Incremental learning based on the concepts of density, typicality and data clouds. • Human interpretable rule-based system. • Assistance in decision making for medical diagnosis. [ABSTRACT FROM AUTHOR]