All issues

The haulage of timber from remote logging areas of Siberia is characterized by high transportation costs. Long stretches of unpaved logging roads, seasonal constraints (spring thaw and winter roads), and considerable distances from cutting areas to main transport hubs (50–200 km) make systematic route planning essential.

This paper proposes a mathematical model of the two-echelon vehicle routing problem (2E-VRP) adapted to timber haulage in the macro-region of the Urals and Siberia. The model incorporates variable transportation costs (truck mileage), fixed costs of using each vehicle ($c_{fix}$ = 30 000 RUB per trip), and vehicle-capacity constraints. The problem is transformed into a QUBO (quadratic unconstrained binary optimization) formulation; capacity constraints are integrated through the penalty coefficient $\lambda_{cap}$ = 500. This representation makes the model compatible with modern specialized solvers, including quantum annealers and digital annealers.

The model is tested on the real geography of the Ural and Siberian Federal Districts: 207 nodes in the logistics network, 65 customer enterprises, and 8 main transport hubs. The proposed three-stage “Feasible-First” algorithm constructs a feasible routing plan in 1.7 s and reduces total transportation costs by 69% relative to a baseline genetic algorithm. Subsequent simulated annealing further improves the QUBO objective value by 55%.

The advent of laser scanning technologies has revolutionized forestry. Their use made it possible to switch from studying woodlands using manual measurements to computer analysis of stereo point images called point clouds.

Automatic calculation of some tree parameters (such as trunk diameter) using a point cloud requires the removal of foliage points. To perform this operation, a preliminary segmentation of the stereo image into the “foliage” and “trunk” classes is required. The solution to this problem often involves the use of machine learning methods.

One of the most popular classifiers used for segmentation of stereo images of trees is a random forest. This classifier is quite demanding on the amount of memory. At the same time, the size of the machine learning model can be critical if it needs to be sent by wire, which is required, for example, when performing distributed learning. In this paper, the goal is to find a classifier that would be less demanding in terms of memory, but at the same time would have comparable segmentation accuracy. The search is performed among classifiers such as logistic regression, naive Bayes classifier, and decision tree. In addition, a method for segmentation refinement performed by a decision tree using logistic regression is being investigated.

The experiments were conducted on data from the collection of the University of Heidelberg. The collection contains hand-marked stereo images of trees of various species, both coniferous and deciduous, typical of the forests of Central Europe.

It has been shown that classification using a decision tree, adjusted using logistic regression, is able to produce a result that is only slightly inferior to the result of a random forest in accuracy, while spending less time and RAM. The difference in balanced accuracy is no more than one percent on all the clouds considered, while the total size and inference time of the decision tree and logistic regression classifiers is an order of magnitude smaller than of the random forest classifier.

The work presents an experimental study of the tree structure that occurs during a medical examination. At each meeting with a medical specialist, the patient receives a certain number of areas for consulting other specialists or for tests. A tree of directions arises, each branch of which the patient should pass. Depending on the branching of the tree, it can be as final — and in this case the examination can be completed — and endless when the patient’s goal cannot be achieved. In the work both experimentally and theoretically studied the critical properties of the transition of the system from the forest of the final trees to the forest endless, depending on the probabilistic characteristics of the tree.

For the description, a model is proposed in which a discrete function of the probability of the number of branches on the node repeats the dynamics of a continuous gaussian distribution. The characteristics of the distribution of the Gauss (mathematical expectation of $x_0$, the average quadratic deviation of $\sigma$) are model parameters. In the selected setting, the task refers to the problems of branching random processes (BRP) in the heterogeneous model of Galton – Watson.

Experimental study is carried out by numerical modeling on the final grilles. A phase diagram was built, the boundaries of areas of various phases are determined. A comparison was made with the phase diagram obtained from theoretical criteria for macrosystems, and an adequate correspondence was established. It is shown that on the final grilles the transition is blurry.

The description of the blurry phase transition was carried out using two approaches. In the first, standard approach, the transition is described using the so-called inclusion function, which makes the meaning of the share of one of the phases in the general set. It was established that such an approach in this system is ineffective, since the found position of the conditional boundary of the blurred transition is determined only by the size of the chosen experimental lattice and does not bear objective meaning.

The second, original approach is proposed, based on the introduction of an parameter of order equal to the reverse average tree height, and the analysis of its behavior. It was established that the dynamics of such an order parameter in the $\sigma = \text{const}$ section with very small differences has the type of distribution of Fermi – Dirac ($\sigma$ performs the same function as the temperature for the distribution of Fermi – Dirac, $x_0$ — energy function). An empirical expression has been selected for the order parameter, an analogue of the chemical potential is introduced and calculated, which makes sense of the characteristic scale of the order parameter — that is, the values of $x_0$, in which the order can be considered a disorder. This criterion is the basis for determining the boundary of the conditional transition in this approach. It was established that this boundary corresponds to the average height of a tree equal to two generations. Based on the found properties, recommendations for medical institutions are proposed to control the provision of limb of the path of patients.

The model discussed and its description using conditionally-infinite trees have applications to many hierarchical systems. These systems include: internet routing networks, bureaucratic networks, trade and logistics networks, citation networks, game strategies, population dynamics problems, and others.

Gesture recognition is an urgent challenge in developing systems of human-machine interfaces. We analyzed machine learning methods for gesture classification based on electromyographic muscle signals to identify the most effective one. Methods such as the naive Bayesian classifier (NBC), logistic regression, decision tree, random forest, gradient boosting, support vector machine (SVM), $k$-nearest neighbor algorithm, and ensembles (NBC and decision tree, NBC and gradient boosting, gradient boosting and decision tree) were considered. Electromyography (EMG) was chosen as a method of obtaining information about gestures. This solution does not require the location of the hand in the field of view of the camera and can be used to recognize finger movements. To test the effectiveness of the selected methods of gesture recognition, a device was developed for recording the EMG signal, which includes three electrodes and an EMG sensor connected to the microcontroller and the power supply. The following gestures were chosen: clenched fist, “thumb up”, “Victory”, squeezing an index finger and waving a hand from right to left. Accuracy, precision, recall and execution time were used to evaluate the effectiveness of classifiers. These parameters were calculated for three options for the location of EMG electrodes on the forearm. According to the test results, the most effective methods are $k$-nearest neighbors’ algorithm, random forest and the ensemble of NBC and gradient boosting, the average accuracy of ensemble for three electrode positions was 81.55%. The position of the electrodes was also determined at which machine learning methods achieve the maximum accuracy. In this position, one of the differential electrodes is located at the intersection of the flexor digitorum profundus and flexor pollicis longus, the second — above the flexor digitorum superficialis.

The aim of study was to develop an ensemble machine learning method for constructing interpretable predictive models and to validate it using the example of predicting in-hospital mortality (IHM) in patients with ST-segment elevation myocardial infarction (STEMI).

A retrospective cohort study was conducted using data from 5446 electronic medical records of STEMI patients who underwent percutaneous coronary intervention (PCI). Patients were divided into two groups: 335 (6.2%) patients who died during hospitalization and 5111 (93.8%) patients with a favourable in-hospital outcome. A pool of potential predictors was formed using statistical methods. Through multimetric categorization (minimizing p-values, maximizing the area under the ROC curve (AUC), and SHAP value analysis), decision trees, and multivariable logistic regression (MLR), predictors were transformed into risk factors for IHM. Predictive models for IHM were developed using MLR, Random Forest Risk Factors (RandFRF), Stochastic Gradient Boosting (XGboost), Random Forest (RF), Adaptive boosting, Gradient Boosting, Light Gradient-Boosting Machine, Categorical Boosting (CatBoost), Explainable Boosting Machine and Stacking methods.

Authors developed the RandFRF method, which integrates the predictive outcomes of modified decision trees, identifies risk factors and ranks them based on their contribution to the risk of adverse outcomes. RandFRF enables the development of predictive models with high discriminative performance (AUC 0.908), comparable to models based on CatBoost and Stacking (AUC 0.904 and 0.908, respectively). In turn, risk factors provide clinicians with information on the patient’s risk group classification and the extent of their impact on the probability of IHM. The risk factors identified by RandFRF can serve not only as rationale for the prediction results but also as a basis for developing more accurate models.

Posts on social networks can both predict the movement of the financial market, and in some cases even determine its direction. The analysis of posts on Twitter contributes to the prediction of cryptocurrency prices. The specificity of the community is represented in a special vocabulary. Thus, slang expressions and abbreviations are used in posts, the presence of which makes it difficult to vectorize text data, as a result of which preprocessing methods such as Stanza lemmatization and the use of regular expressions are considered. This paper describes created simplest machine learning models, which may work despite such problems as lack of data and short prediction timeframe. A word is considered as an element of a binary vector of a data unit in the course of the problem of binary classification solving. Basic words are determined according to the frequency analysis of mentions of a word. The markup is based on Binance candlesticks with variable parameters for a more accurate description of the trend of price changes. The paper introduces metrics that reflect the distribution of words depending on their belonging to a positive or negative classes. To solve the classification problem, we used a dense model with parameters selected by Keras Tuner, logistic regression, a random forest classifier, a naive Bayesian classifier capable of working with a small sample, which is very important for our task, and the k-nearest neighbors method. The constructed models were compared based on the accuracy metric of the predicted labels. During the investigation we recognized that the best approach is to use models which predict price movements of a single coin. Our model deals with posts that mention LUNA project, which no longer exist. This approach to solving binary classification of text data is widely used to predict the price of an asset, the trend of its movement, which is often used in automated trading.