Publications

Many economic applications including optimal pricing and inventory management requires prediction of demand based on sales data and estimation of sales reaction to a price change. There is a wide range of econometric approaches which are used to correct a bias in estimates of demand parameters on censored sales data. These approaches can also be applied to various classes of machine learning models to reduce the prediction error of sales volume. In this study we construct two ensemble models for demand prediction with and without accounting for demand censorship. Accounting for sales censorship is based on the idea of censored quantile regression method where the model estimation is splitted on two separate parts: a) prediction of zero sales by classification model; and b) prediction of non-zero sales by regression model. Models with and without accounting for censorship are based on the predictions aggregations of Least squares, Ridge and Lasso regressions and Random Forest model. Having estimated the predictive properties of both models, we empirically test the best predictive power of the model that takes into account the censored nature of demand. We also show that machine learning method with censorship accounting provide bias corrected estimates of demand sensitivity for price change similar to econometric models.

Many economic applications including optimal pricing and inventory management requires prediction of demand based on sales data and estimation of sales reaction to a price change. There is a wide range of econometric approaches which are used to correct a bias in estimates of demand parameters on censored sales data. These approaches can also be applied to various classes of machine learning models to reduce the prediction error of sales volume. In this study we construct two ensemble models for demand prediction with and without accounting for demand censorship. Accounting for sales censorship is based on the idea of censored quantile regression method where the model estimation is splitted on two separate parts: a) prediction of zero sales by classification model; and b) prediction of non-zero sales by regression model. Models with and without accounting for censorship are based on the predictions aggregations of Least squares, Ridge and Lasso regressions and Random Forest model. Having estimated the predictive properties of both models, we empirically test the best predictive power of the model that takes into account the censored nature of demand. We also show that machine learning method with censorship accounting provide bias corrected estimates of demand sensitivity for price change similar to econometric models.

With an increased interest in machine processable data and with the progress of semantic technologies, many datasets are now published in the form of RDF triples for constituting the so-called Web of Data. Data can be queried using SPARQL but there are still needs for integrating, classifying and exploring the data for data analysis and knowledge discovery purposes. This research work proposes a new approach based on Formal Concept Analysis and Pattern Structures for building a pattern concept lattice from a set of RDF triples. This lattice can be used for data exploration and in particular visualized thanks to an adapted tool. The specific pattern structure introduced for RDF data allows to make a bridge with other studies on the use of structured attribute sets when building concept lattices. Our approach is experimentally validated on the classification of RDF data showing the efficiency of the underlying algorithms.

Studying the heterogeneity of consumers allows to price the product differently for consumer segments or groups of a product. In this paper we estimate a model of aggregate demand for Perm Opera and Ballet Theatre focusing on the heterogeneity in price effect on demand for tickets on different performances and seats. We estimate parameters of demand function using censored quantile regression that accounts for the limited capacity of the theatre house. We reveal the price effect variation across different types of theatrical productions and seats with lower elastic demand on ballets and for seats of higher quality.

  In many practical tasks it is needed to estimates effect of a treatemnt on individual level. For example in medicine it is essential to determine the patients that would benifit from a certain medicament. In marketing knowning the persons that are likely to buy a new product would reduce the amount of spam. In this chapter we review the methods to estimate individulize treatment effect from a randomized trial, i.e., an experiment when a part of individuals recieves a new treatment, while the others does not. Finally, it is shown that new efficient methods are needed in this domain.

In this paper, we analyze a new approach for demand prediction in retail. One of the signicant gaps in demand prediction by machine learning methods is the unaccounted sales data censorship. Econometric approaches to modeling censored demand are used to obtain consistent and unbiased estimates of parameters. These approaches can also be transferred to different classes of machine learning models to reduce the prediction error of sales volume. In this study we build two ensemble models to predict demand with and without demand censorship, aggregating predictions for machine learning methods such as Linear regression, Ridge regression, LASSO and Random forest. Having estimated the predictive properties of both models, we test the best predictive power of the models with accounting for the censored nature of demand.

Government-subsidized theatres in Russia aim to increase both revenue and number of tickets sold. Achievement of such a goal is closely connected with study of theatre tickets demand function focusing on the estimation of price elasticity of demand. Previous papers studied price elasticity of demand for theatrical performances provide controversial results. Estimates of price elasticity varies with countries, theatres, segments of theatre audience and even with level of data aggregation. Ticket sales and price data aggregation at the national level or theatre level typically does not allow to control for the differences in produced cultural goods quality. Incorrect control or ignoring of a theatre, performances and seats quality leads to a problem of omitted variables and bias in price elasticity estimates. In this paper we employ disaggregated ticket sales data for four seasons of Perm opera and ballet theatre and estimate the price effect on the theatrical demand for various performance types and seats in the hall. We use censored quantile regression to estimate the parameters of theatre demand function. We reveal the weak elasticity of demand for Perm opera and ballet theatre. We also show that ignoring the limited capacity of the theatre hall, characteristics of performances and seats quality leads to biased estimates of demand function parameters.

Russian book market is one of the largest in the world in terms of new titles in print. However, this market is  underexplored. There is no research dealing with an empirical demand or supply function estimation for this market. The purpose of this paper is to analyze the book demand function and to check whether this kind of demand is price and income elastic. On the basis of results retrieved, managerial recommendations are to be offered. For this purpose, the demand function for books is built and estimated both on the total sample and for particular literature genres. The peculiarity of the demand model estimation is the introduction in the model covariates indicating the book content quality such as the amount of people who gave rating on the website, average rating of the book from the website and the combined effect of these two variables. An empirical estimation of these factors influence has not been considered in the previous research yet. Model estimation was based on the data of one North-Western Federal District book retail chain. According to the estimation results, book demand is price inelastic; moreover, books are estimated to be luxury goods. The analysis of demand functions for separate genres suggests that the demand for each genre is price-insensitive. Only Russian and foreign prose, foreign fiction and poetry are luxury goods among all the genres analyzed. A foreign detective, Russian fiction and sentimental novel are normal goods, whereas Russian detective is an inferior goods. The results of the research might be of a particular interest for books retail chains and publishers

Recovery after stroke relates tightly to the white matter integrity. Currently, the main methodology for non-invasive white matter integrity assessment is diffusion-weighted magnetic resonance imaging (DW-MRI), a state-of-the-art approach which is, however, prone to multiple limitations. Using DW-MRI, it was demonstrated that many pathways including corticospinal tract (CST) and corpus callosum contribute to structural brain reserve after stroke, but only a few of these tracts were found to be useful in the clinical practice. The most widely known measure is an asymmetry of the fractional anisotropy (FA) in CST at the level of the internal capsule, which could be used for predicting motor recovery in acute stroke. Recently, a new complementary motor component of the structural reserve, the so-called alternate motor fibers (AMFs), was proposed for motor recovery prognosis in stroke patients, and it was even reported to correlate with the effect of the transcranial direct current stimulation in chronic stroke. Here, we would like to point out a possible additional sensory interpretation of the AMF that appears plausible after taking into account technical limitations of DW-MRI approach, which may potentially give rise to different interpretations of the same results.

In the times of Soviet Union books were a luxury good. This paper examines whether books are still a luxury good in Russia. For this purpose data from one of the Russian book retail chains is used to empirically estimate a general book demand and separate demand models for genres. We focus on estimating income elasticity. For this reason we construct a covariate on the basis of monthly wages of working individuals that reveals consumer income. Moreover, this paper is one of the few which addresses in detail the influence of books content quality on book demand. The main result is that books on average are not luxury goods anymore in Russia. However two genres: foreign prose and poetry are exceptions and can be called luxury goods. We also conclude that quality control covariates (book rating and number of people who rated the book) are important determinants of book demand as they influence significantly the general book demand and the demand models for different genres as well.

In this study, we use a sample of 192 listed shipping companies and employ a logit model in order to investigate the determinants of the probability of default. We enhance our analysis by isolating not only the cases of company liquidations but also those cases where companies had to change their legal status due to warning liquidity signals. Our key findings are in line with prior research and moreover we depict a changing trend in the marginal effects of relevant variables, on the probability of default. We further show, through an empirical application, how the obtained results can be used in a managerial decision-making process and in a bank credit underwriting process in order to assess the creditworthiness of a shipping company.

A scalable method for mining graph patterns stable under subsampling is proposed. The existing subsample stability and robustness measures are not antimonotonic according to definitions known so far. We study a broader notion of antimonotonicity for graph patterns, so that measures of subsample stability become antimonotonic. Then we propose gSOFIA for mining the most subsample-stable graph patterns. The experiments on numerous graph datasets show that gSOFIA is very efficient for discovering subsample-stable graph patterns.  

People are intent to make similar choices especially in consumer goods markets. To address both explanations of this persistence, i.e. state dependence and heterogeneity in preferences, we use random coefficient logit model based on scanner panel data on juice purchases. The product differentiation of the chosen category allows us to model three dimensions of state dependence on brand, size and flavor characteristics. We provide evidence that the persistence in brand choices is positively correlated with persistence in size and flavor choices, thus the consumer pattern is prone to be inertial or variety seeking in every product characteristics. Simultaneously we show that the more sensitive to price and promotional activities consumers are, the less inertial is their behavior. 

The present study analyzes Perm, Russia residential housing market supply focusing on sellers' heterogeneity. Many indicators of heterogeneity were consi- dered in the previous research, and all of them were proved to have a great impact on housing prices and time on the market. However, the gap exists in evaluating sellers’ pricing strategies in dynamics mostly because of unavailable data. Current study clears out the effect of time on price using data on asking price dynamics. We employ semiparametric sample selection estimation proce- dure which accounts for the unobserved property characteristics and non-random selection of objects out of the sample. We consider two main types of sellers: real estate agents and property owners, and show that real estate agents appear to be more impatient compared to property owners. Specifically, they set a lower asking price initially and are more likely to revise it over time if the object is not sold.

Purpose: This paper examines heterogeneity of preferences of mortgage borrowers of Russian state-owned supplier of residential housing mortgages. Methodology: Analysis takes into account the underwriting process and the choice of contract terms of all loans originated from 2008 to 2012. Our dataset contains demographic and financial characteristics for all applications, loan terms and the performance information for all issued loans by one regional bank which operates government mortgage programs. We use a multistep semiparametric approach to estimate the determinants of bank and borrower choice controlling for possible heterogeneity of preferences, sample selection and endogeneity of contract terms. Findings: We found that demand of low-income households who are unable to afford improving of housing conditions by other instruments than government mortgage is less elastic according to the change both in interest rate and maturity compared with higher incomed households. Originality: The main contribution to the literature is modeling choice of contract terms as interdependent by structural system of simultaneous equations with heterogeneous marginal effects.

Symbolic classifiers allow for solving classification task and provide the reason for the classifier decision. Such classifiers were studied by a large number of researchers and known under a number of names including tests, JSM-hypotheses, version spaces, emerging patterns, proper predictors of a target class, representative sets etc. Here we consider such classifiers with restriction on counter-examples and discuss them in terms of pattern structures. We show how such classifiers are related. In particular, we discuss the equivalence between good maximally redundant tests and minimal JSM-hyposethes and between minimal representations of version spaces and good irredundant tests.

In this research we analyze new approach for prediction of demand. In the studied market of  performing arts the observed demand is limited by capacity of the house. Then one needs to account for demand censorhip to obtain unbiased estimates of demand funnction parameters. The presence of consumer segments with dierent purposes of going to the theatre and willingness-to-pay for performance and ticket characteristics causes a heterogeneity in theatre demand. We propose an estimator for prediction of demand that accounts for both demand censorhip and preferences heterogeneity. The estimator is based on the idea of classiffication and regression trees and bagging prediction aggregation extended for prediction of censored data. Our algorithm predicts and combines predictions for both discrete and continuous parts of censored data.We show that our estimator performs better in terms of prediction accuracy compared with estimators which accounts either for censorship, or heterogeneity only. The proposed approach is helpful for finding product segments and optimal price setting.

Purpose – This study explores company strategies for intangibles. We investigate whether it is reasonable for companies to intensify intangibles when the current strategy is not intangible-intensive. This paper aims to elaborate a theoretical model to describe the strategic decision-making in companies.

Design/methodology/approach – We use the Bellman equation framework to find the conditions under which a change in strategy for intangibles is reasonable.

Findings – The results determine the parameters of returns on intangibles in different strategies, the optimal intangible stock and the influence of external economic shocks. The findings of our study demonstrate that many requirements have to be met to make intangible-intensive strategy beneficial for a company. Moreover negative shocks of crises force a company to postpone a new strategy on intangibles.

Practical implications – This research provides an insight into strategic behaviour of companies under uncertainty. The theoretical findings demonstrate under which conditions companies should decide to switch to a strategy more intangible-intensive. This model can be used to empirically test parameters of different investment strategies of companies using structural estimation techniques.

Originality/value – This work contributes to the theory of managerial economics giving closed form solutions for the dynamic optimization of company behaviour. The findings also show how this behaviour might change when economic crises are faced or expected.