Using a Genetic Algorithm to Derive a Linguistic Summary of Trends in Numerical Time Series

Abstract

The purpose of this paper is to propose a new easily implementable approach to a linguistic summarization of trends that may occur in temporal data, to be more specific -time series. To characterize the trends in time series, we use three parameters: dynamics of change, duration and variability, and apply to them the fuzzy linguistic summaries of data (databases) in the sense of Yager (cf. Yager [13], Kacprzyk and Yager [7] and Kacprzyk, Yager and Zadrozny [8]) which in the form of natural language-like sentences subsume the very essence of a set of data. A genetic algorithm is used to generate the linguistic summaries sought.

Linguistic Summarization of Europe Brent Spot Price Time Series Along with the Interpretations from the Perspective of Turkey

“…Time was usually partitioned into groups by hour, day, or month. Kacprzyk et al [19][20][21][22][23][24][25][26][27]39,41,[74][75][76][77][78][79] first identified local (partial) trends of time series and then considered three aspects of local trends for summarization of time series. The behaviors of time series were obtained by partitioning the time, and the attributes into each segments of which were labeled with fuzzy sets.…”

“…There have been a numerous studies reported on linguistic summarization based upon fuzzy sets with different approaches applied to various areas: linguistic summarization of time series, decision support systems, text categorization, reconciliation processes, human motion analysis, social networks, monitoring of internet traffic flows, fall detection, financial reports, driving and the traffic activities . Undoubtedly, one of the most important application areas of linguistic summarization is time series data mining (TSDM) with generating propositions describing trends of time series in terms of the following three aspects: duration, variability, and dynamics of change . For instance, “trends with low variability that took most of the time” is a proposition that can be obtained by linguistic summarization.…”

“…34,35 Undoubtedly, one of the most important application areas of linguistic summarization is time series data mining (TSDM) with generating propositions describing trends of time series in terms of the following three aspects: duration, variability, and dynamics of change. [20][21][22][36][37][38][39][40][41] For instance, "trends with low variability that took most of the time" is a proposition that can be obtained by linguistic summarization. It is obvious that propositions like that are more convenient to make future plans and more useable for decision makers.…”

Extracting an Optimal Set of Linguistic Summaries using Genetic Algorithm Combined with Greedy Strategy

“…However, human users can discover interesting relationships and knowledge hidden in the dataset. The studies in [8,[15][16][17][18][19] applied genetic algorithm to find an optimal set of linguistic summaries. Therefore, human users need to define the constraints and a quality evaluation function for the set of linguistic summaries based on the user's needs.…”