The indoor positioning task is an exciting problem because of the practical applications and the scientific challenges. This work emphasises the impact of signal source selection on the effectiveness of positioning models relying on Wi-Fi received signal strength data by eliminating unstable signal sources such as mobile access points. Two localisation algorithms, based on k-Nearest Neighbour and Random Forest, were applied to data from three academic buildings to test the effectiveness of various access points (APs) selection filters. Additionally, new hierarchical selection models were proposed. The methods divide building into zones to improve the characteristics of the positioning error by a local selection of APs. Hierarchical methods reduced the median number of AP to 67 in comparison to over 100 selected by existing methods, obtaining a statistically similar median error of 6.01 m.

Spatial data on a cellular network load can be used to develop commercial and public services. However, such data is calculated based on individual users’ behavior and can contravene their privacy rights. Moreover, direct tracking of individual devices violates the European Union’s regulations. To solve this issue, we propose to use data aggregated in individual cells of the public land mobile network without tracking an individual mobile device in the entire process. To prove that the proposed data collection method is useful, we compared the obtained results with a closed-circuit television system in an estimation of the number of people. The proposed system is sensitive enough to detect untypical global events in an urban area and distinguish transport demand zones of various types as we showed on real data from the City of Warsaw.

This paper presents an algorithm predicting oil level and temperature sensor (OLTS) failure to replace it before it carries serious costs. OLTS sensor showing too high oil temperature cockpit indications is a driver of significant air turnback events and commanded in-flight shutdown (IFSD). A prediction of sensor malfunction is possible, but an operator requires at least 11 months of historical data. The developed algorithm automates the process of identifying potential failures using a data-driven, dissimilarity based model. It calculates the rolling mean of the oil temperature difference between sister engines for short-term and long-term periods (counted in flights). If the difference between the short-term and long-term means is greater than a set threshold at least confirmation window times, it sets an alert. The proposed model requires less than three months of data to detect the malfunction, with the final F1 score measured on the test set equal to 0.71.

In this work, we describe an urban Internet of Things (IoT) architecture, grounded in big data patterns and focused on the needs of cities and their key stakeholders. First, the architecture of the dedicated platform USE4IoT (Urban Service Environment for the Internet of Things), which gathers and processes urban big data and extends the Lambda architecture, is proposed. We describe how the platform was used to make IoT an enabling technology for intelligent transport planning. Moreover, key data processing components vital to provide high-quality IoT data streams in a near-real-time manner are defined. Furthermore, tests showing how the IoT platform described in this study provides a low-latency analytical environment for smart cities are included.

Machine learning techniques are a standard approach in spam detection. Their quality depends on the quality of the learning set, and when the set is out of date, the quality of classification falls rapidly. The most popular public web spam dataset that can be used to train a spam detector—WEBSPAM-UK2007—is over ten years old. Therefore, there is a place for a lifelong machine learning system that can replace the detectors based on a static learning set. In this paper, we propose a novel web spam recognition system. The system automatically rebuilds the learning set to avoid classification based on outdated data. Using a built-in automatic selection of the active classifier the system very quickly attains productive accuracy despite a limited learning set. Moreover, the system automatically rebuilds the learning set using external data from spam traps and popular web services. A test on real data from Quora, Reddit, and Stack Overflow proved the high recognition quality. Both the obtained average accuracy and the F-measure were 0.98 and 0.96 for semiautomatic and full–automatic mode, respectively.

This work describes the whole process of 3D model reconstruction.It begins with the representation of the method that is used to find the matching between photographs and the methodology to use the data to form the initial structure of the reconstructed model, represented by a point cloud.As a next stage, a refinement process is performed, using the bundle adjustment method.A set of stereovision methods is used later on to find a more detailed solution.Those algorithms use pairs of images, therefore as a prerequisite a set of routines that aggregates those results is studied.The paper is concluded with a description of how the point cloud is processed, including the surface reconstruction, to form the result.The described methodology is illustrated with reconstructions of three series of professional photographs from a public repository and one series of amateur photographs created especially for this work.The results were evaluated by the proposed area matching and contour matching measures.

The task of identifying native and foreign elements and rejecting foreign ones in the pattern recognition problem is discussed in this paper. Such the task is a nonstandard aspect of pattern recognition, which is rarely present in research. In this paper, ensembles of support vector machines solving two-classes and one-class problems are employed as classification tools and as basic tools for rejecting of foreign elements. Evaluation of quality of classification and rejection methods are proposed in the paper and finally some experiments are performed in order to illustrate acquainted terms and methods.

This paper describes a computer vision based automatic scoring system of shooting targets. The system estimates scoring with a professional tournament precision, but is dedicated to amateur shooters and can work with photos taken by amateur cameras and mobile devices. The automatic scoring issue is divided into three problems: a target detection, a holes detection, and a hole analysis. The target is detected on the base of a bull–eye localization. The holes detection bases on the Hough transformation. The holes analysis localizes a position of hole’s center. The position relative to detected scoring sections is a base for scoring. The proposed algorithm detects holes with 99 percent accuracy. An elimination of false positives results reduces the level of accepted holes to 92 percents. The average error for the automatic score estimation is 0.05 points. The estimation error for over 91 percent holes is lesser than a tournament–scoring threshold.

This paper presents a pattern recognition study aimed al music symbols recognition. The study is focused on classification methods of music symbols based on decision trees and clustering method applied to classes of music symbols that face classification problems. Classification is made on the basis of extracted features. A comparison of selected classifiers was made on some classes of nutation symbols distorted by a variety of factors as image noise, printing defects, different fonts, skew and curvature of scanning, overlapped symbols.