The study’s premise was to analyze and predict property prices in the US local market, taking into account various factors that influence property values. The aim was to compare and evaluate machine learning regression models to estimate housing and house prices based on the data available for a given geographical area (the US). This study analyses the relationship between attributes describing the US property market and the price of that property and then selects them accordingly. The selection aims to identify variables whose elimination will improve the properties of the regression model we wish to build. Regularisation is applied, a technique used in machine learning to prevent models from becoming too complex, which can lead to overfitting. Two regularisation techniques were proposed: LASSO and ElasticNet. Reports available from the Joint Center for Housing Studies of Harvard University have shown that many criteria affect the price of housing. On this basis, this paper proposes extensions to the model. The results of the analyses of the LASSO and ElasticNet models indicate the superior performance of LASSO in various data analysis scenarios, especially in the context of spatial and geolocation data.

The article is devoted to the selection of parameter values of the evolutionary algorithm for creating new images. Based on simple geometric figures, an impression of them is created using the classical approach of the evolutionary algorithm, and these obtained results are subjected to genotypic (a measure of non-overlap with the edge matrix) and phenotypic evaluation: using for this purpose measures of aesthetics formed on the basis of Benford’s law and fractal dimension. The aim of the paper is to find optimal values of the algorithm’s parameters so that the aesthetics evaluation measures are satisfactory. The paper also shows the detected relationships by analyzing Pearson correlation measures.

Many recommendation algorithms have been proposed in the literature that generates personalized recommendations. These recommendations are often presented to the user as an ordered list of suggested items (so-called top-N recommendation). However, despite years of research, no algorithm has been proposed to generate high-quality recommendations for all users in the system. To improve the quality of the final recommendation, aggregation techniques may be employed to combine the results returned by several recommendation algorithms. However, prior to the aggregation process, various normalization techniques are often utilized. This paper will present the results of experiments designed to investigate how different normalization strategies affect the quality of the created aggregation. The research was conducted using four such strategies and ten unsupervised aggregation methods on the publicly available MovieLens 100k dataset. Results were validated by statistical tests.

The paper presents the adaptation process of an Ant Colony Optimization system to the task of composing music. The main aims of the study include designing and implementing the above-mentioned system, conducting research, measuring the obtained results, and drawing conclusions. The application was implemented in Python and it allows you, based on the loaded MIDI file, to generate new melodies played on various instruments. The way of composing a melody is configurable by modifying the parameters of the ACO algorithm and different composition modes. To measure the quality of the obtained results, an evaluation function was created, which is used in research to measure the correlation between the parameters of the system. Based on the conducted research, it was determined how individual parameters affect the results obtained by the system. The paper covers the presentation of theoretical knowledge, system design, implementation description, research, and conclusion.

Over the years, various techniques of generating recommendations have been developed. However, it turns out that when we compare the recommendations generated by different algorithms in the context of a particular user, the quality of such recommendations for different techniques may differ. The use of the aggregation techniques, the aim of which is to combine several rankings into one, can be a solution to this problem. In theory it should improve the quality of the recommendations. Additionally, in order to personalize the recommendations better, a metaheuristic algorithm, which, by assigning different weights to each feature, tries to represent the preference of the active user, was used. This paper also presents a suggestion to include additional rankings generated for other users in the system in the aggregation process. The idea will be supported by research results that clearly show that taking into account rankings of other users can improve the quality of the generated recommendations.

Clustering analysis is used in many disciplines and applications. This data mining task is an essential tool that identifies groups of objects based on similarity measures. The ant clustering algorithm is a swarm-intelligent method used for clustering problems, that is inspired by the natural behavior of ants, which collect their corpses and sort their larvae. In recent years, many new metaheuristic algorithms have explored different scientific fields using inspirations or motivations in biological or nature-inspired derivations. The same situation is observed in area of data mining, where plenty of metaheuristics was presented as a new and effective approach. This state causes some difficulties or disappointments to have enough explanations or deep extent analysis to justify such applications. A new version of an ant clustering algorithm, using cellular automata mechanisms of transitions, is proposed to examine the computational efficiency and accuracy of the ant clustering approach. Based on the similarity of two approaches arising from Ant Sleeping Model, computational analysis of difficulties and complexities are presented and results show that the modification of an ant clustering algorithm - ASMMOD produces results that are not only more stable but also more efficiently determined than the prototype of ASM. The aim of this paper is not explicitly criticize these approaches but also to find a way of the repair these algorithms.

Automated teller machines are affected by two kinds of attacks: physical and logical. It is common for most banks to look for zero-day protection for their devices. The most secure solutions available are based on complex security policies that are extremely hard to configure. The goal of this article is to present a concept of using the modified MajorClust algorithm for generating a sandbox-based security policy based on ATM usage data. The results obtained from the research prove the effectiveness of the used techniques and confirm that it is possible to create a division into sandboxes in an automated way.

Recommendation systems can suggest users list of items they have not yet seen but might be interested in. To improve the quality of the generated recommendations, different techniques are often used which try to personalize recommendations. Usually user preferences are stored in the form of a vector in which individual values describe to what extent a given feature is desired by the user. To find this vector, metaheuristic algorithms can be used, however their main drawback is their computational complexity. Therefore, in this paper, a modification of the Differential Evolution algorithm is proposed to enable faster computation of the ranking score for each item in the system, which is used to create a recommendation list. Experiments have been performed on the current MovieLens 25m database and they show that our modification can significantly speed up the process of finding a preference vector, without losing their quality for the top-N recommendation task. We will also address the vulnerability of recommendation systems to profile injection attacks, as a result of which an attacker can influence the generated recommendations.

Abstract The article discusses the utilitarian problem of the mobile collection of waste electrical and electronic equipment. Due to its $$\mathcal {NP}$$ NP -hard nature, implies the application of approximate methods to discover suboptimal solutions in an acceptable time. The paper presents the proposal of a novel method of designing the Evolutionary and Memetic Algorithms, which determine favorable route plans. The recommended methods are determined using quality evaluation indicators for the techniques applied herein, subject to the limits characterizing the given company. The proposed Memetic Algorithm with Tabu Search provides much better results than the metaheuristics described in the available literature.

The recommendation systems are used to suggest new, still not discovered items to users. At the moment, in order to achieve the best quality of the generated recommendations, users and their choices in the system must be analyzed to create a certain profile of preferences for a given user in order to adjust the generated recommendation to his personal taste. This article will present a recommendation system, which based on the Differential Evolution (DE) algorithm will learn the ranking function while directly optimizing the average precision (AP) for the selected user in the system. To achieve that, items are represented through a feature vectors generated using user-item matrix factorization. The experiments have been conducted on a popular and widely available public dataset MovieLens, and show that our approach in certain situations can significantly improve the quality of the generated recommendations. Results of experiments are compared with other techniques.

Automated teller machines are affected by two kinds of attacks: physical attacks, and software-based attacks, whereas the latter is becoming more and more popular every day. Most banks tend to look for a day-zero protection in order to secure their devices. Two most popular mechanisms are whitelisting and sandboxing. However, this kind of protection is not only hard to configure, but it also requires a deeper knowledge about software security as well as the information about the software currently installed on the device. The goal of this article is to present a possibility of using a modified DBSCAN algorithm to solve a complex configuration problem which is clustering programs within operating system. Results of the experimental studies show that this kind of automatic configuration can be even more precise than the basic clustering algorithms. As the results show great promise, it is all right to believe that the algorithms could be used in security products if additional security rules are taken into account.

This paper presents a detailed study of the discrete particle swarm optimization algorithm (DPSO) applied to solve the dynamic traveling salesman problem which has many practical applications in planning, logistics and chip manufacturing. The dynamic version is especially important in practical applications in which new circumstances, e.g., a traffic jam or a machine failure, could force changes to the problem specification. The DPSO algorithm was enriched with a pheromone memory which is used to guide the search process similarly to the ant colony optimization algorithm. The paper extends our previous work on the DPSO algorithm in various ways. Firstly, the performance of the algorithm is thoroughly tested on a set of newly generated DTSP instances which differ in the number and the size of the changes. Secondly, the impact of the pheromone memory on the convergence of the DPSO is investigated and compared with the version without a pheromone memory. Moreover, the results are compared with two ant colony optimization algorithms, namely the $$\mathcal {MAX}$$ – $$\mathcal {MIN}$$ ant system (MMAS) and the population-based ant colony optimization (PACO). The results show that the DPSO is able to find high-quality solutions to the DTSP and its performance is competitive with the performance of the MMAS and the PACO algorithms. Moreover, the pheromone memory has a positive impact on the convergence of the algorithm, especially in the face of dynamic changes to the problem’s definition.

The number of highly sophisticated software-based attacks on ATMs is growing these days. New types of threats require new ways of system protection. Most secure solutions are based on whitelists and sandboxes, which unlike antivirus solutions are able to protect the system against any new threat. Sadly, they are hard to configure therefore require an expert-level knowledge of operating system mechanisms and software security techniques. The main purpose of this article is to present the possibilities of using clustering algorithms in a configuration process of a sandbox-based security solution. Results of the experimental studies show that even basic clustering algorithms can be used as a part of the configuration process. Achieved results were deemed promising by the control algorithm and thus can be used as a base for future research.