Driven by the variety of available measures intended to estimate predictability of diverse objects such as time series and network links, this paper presents a comprehensive overview of the existing literature in this domain. Our overview delves into predictability from two distinct perspectives: the intrinsic predictability, which represents a data property independent of the chosen forecasting model and serves as the highest achievable forecasting quality level, and the realized predictability, which represents a chosen quality metric for a specific pair of data and model. The reviewed measures are used to assess predictability across different objects, starting from time series (univariate, multivariate, and categorical) to network links. Through experiments, we establish a noticeable relationship between measures of realized and intrinsic predictability in both generated and real-world time series data (with the correlation coefficient being statistically significant at a 5% significance level). The discovered correlation in this research holds significant value for tasks related to evaluating time series complexity and their potential to be accurately predicted.
1 January 2023 in Journal of Advances in Information Technology
Dynamic pricing is frequently used in online marketplaces, ticket sales, and booking systems. The commercial principles of dynamic pricing systems are often kept secret; however, their application causes complex changes in human behavior. Thus, a scientific tool is needed to evaluate and predict the impact of dynamic pricing strategies. Publications in the field lack a common quality evaluation methodology, public data, and source code, making them difficult to reproduce. In this paper, a data-driven method, DPRank, for evaluating dynamic pricing systems is proposed. DPRank first builds a surrogate price elasticity of demand model using public data generated by a hidden dynamic pricing model, and then applies the surrogate model to build an exposed dynamic pricing model. The hidden and exposed dynamic pricing models were then systematically compared in terms of quality using a Monte Carlo simulation in terms of a company’s revenue. The effectiveness of the proposed method was tested on the dataset collected from the website of a Russian railway passenger carrier company. Depending on the train type, the quality difference between the hidden and exposed models can vary by several dozen percent on average, indicating the potential for improving the existing (hidden) company’s dynamic pricing model.