On some types of linguistic summaries of time series

Kacprzyk, Janusz; Wilbik, Anna; Zadrożny, Sławomir

doi:10.1109/is.2006.348448

Cited by 24 publications

(11 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The details of the algorithm are presented in our previous papers -cf. Kacprzyk, Wilbik, Zadrożny [9,6,5] The resulting ε-approximation of a group of points p_0, . .…”

Section: Temporal Data and Trend Analysismentioning

confidence: 99%

“…In our further papers (cf. Kacprzyk, Wilbik and Zadrożny [9,6,5]) we proposed, first, another type of summaries that uses a linguistic quantifier based aggregation not over the number of trends but over the time instants they take altogether. For example, such a summary can be: "Trends taking most of the time are increasing" or "Increasing trends taking most of the time are of a low variability".…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Analysis of Time Series via their Linguistic Summarization: the Use of the Sugeno Integral

Kacprzyk¹,

Wilbik²,

Zadrożny³

2007

Seventh International Conference on Intelligent Systems Design and Applications (ISDA 2007)

Self Cite

View full text Add to dashboard Cite

We propose here some new types of linguistic summaries of time series by extending our previous works. First, the linguistic summaries of time series refer to the summaries of (partial) trends identified here with straight line segments of a piece-wise linear approximation of a time series that is proposed in the paper. To characterize the trends we use the slope of the line segment, the goodness of approximation and the length of the trend. A linguistic summary of a time series is then derived by a linguistic quantifier driven aggregation of trends that is performed using the Sugeno integral. We show an application to the absolute performance type analysis of time series data on daily quotations of an investment fund over an eight year period. The results are very promising.

show abstract

“…The details of the algorithm are presented in our previous papers -cf. Kacprzyk, Wilbik, Zadrożny [9,6,5] The resulting ε-approximation of a group of points p_0, . .…”

Section: Temporal Data and Trend Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Analysis of Time Series via their Linguistic Summarization: the Use of the Sugeno Integral

Kacprzyk¹,

Wilbik²,

Zadrożny³

2007

Seventh International Conference on Intelligent Systems Design and Applications (ISDA 2007)

Self Cite

View full text Add to dashboard Cite

show abstract

“…In comparison to existing methods, more accurate forecasting methods can be obtained using a rule-based forecasting (RBF), a technique combining data extrapolation [7,13,14,25,26,[43][44][45], time series [28,29,44,45], and elements of expert systems [5-7, 22, 23, 34, 37, 46]. The four most important methods of extrapolation were used: linear regression, random walk, and Brown's exponential smoothing, as well as Holt's exponential smoothing.…”

Section: Introductionmentioning

confidence: 99%

Detecting Nasdaq Composite Index Trends with OFNs

Zarzycki

Czerniak

Dobrosielski

2017

Theory and Applications of Ordered Fuzzy Numbers

View full text Add to dashboard Cite

The chapter presents a novel way of describing changes in the stock index and the identification of potential trends. The authors already used a similar approach to describe the stock exchange index [16]; this chapter is a continuation and another application of work on this issue. The method for detecting patterns in a trend by means of linguistic variables is described. The use of computational operations on numbers in the Ordered Fuzzy Number (OFN) notation [40][41][42] enables us to set the values of linguistic variables and thus conduct fuzzification of the input. By using one OFN number it is possible to store five parameters of index quotations (open, high, low, and close values as well as a change direction). The OFN numbers are conveyed into a linguistic form. In order to find trend sequence similarity the following applies: sequence identity with the input frame expressed as a percentage, frame size, the level of threshold conformity with the frame (threshold), and how often the pattern is present (frequency). A dedicated computer program to detect patterns is implemented. The program used data from the index Nasdaq Composite from the years 2006-2016. The results represent a further step to develop effective methods of rule-based forecasting.

show abstract

“…Particle swarm optimization was created thanks to studies on, among others, sandblasting of a car body or other corroded metal parts. Hence, generally, this branch of AI has been called swarm intelligence [11,14,25,38]. Conversion of those intelligence mechanisms prevailing among simple individuals into the field of computer systems resulted in creation of the current sometimes called computational swarm intelligence.…”

Section: Introductionmentioning

confidence: 99%

OFNAnt Method Based on TSP Ant Colony Optimization

Czerniak

2017

Theory and Applications of Ordered Fuzzy Numbers

View full text Add to dashboard Cite

This chapter presents a hybrid method of swarm intelligence current. Intelligence represented by ant colonies has been enriched with fuzzy logic arithmetics. In this case Kosiński's Ordered Fuzzy Numbers were specifically used. Apart from a fuzzy decision model of a single ant used earlier by other researchers, the author used the order as a trend support. By associating the direction of a number in Ordered Fuzzy Numbers (OFNs) with the trend observed in the ant colony it is possible to provide a unique description of a fuzzy observation of a colony behavior. The experiments were carried out in the area of searching for the optimal connecting route in the field. The experiment covered 10 complex issues of searching for the optimal route. All are benchmarks from the TSPlib repository which are well known among researchers. They represent the actual problems of route selection such as transport connections depending on geographic conditions and optimizing the machining process or the layout of the power networks. The complexity level of optimal solutions for problems to be solved amounted from several hundred to several thousand connections. Each of them was solved using six swarm intelligence methods and five well-known classical methods dedicated to the traveling salesman problem (TSP). The results were presented in the form of tables and graphs, and some of the routes were shown in graphical form. Final conclusions of the experiment indicate the superiority of methods based on ant colony optimization as regards closeness to optimal solutions. The results achieved by the OFNAnt method are generally better (in 92% of cases) than those achieved by classic methods and are in the forefront of solutions from the swarm intelligence group.

show abstract

On some types of linguistic summaries of time series

Cited by 24 publications

References 19 publications

Analysis of Time Series via their Linguistic Summarization: the Use of the Sugeno Integral

Analysis of Time Series via their Linguistic Summarization: the Use of the Sugeno Integral

Detecting Nasdaq Composite Index Trends with OFNs

OFNAnt Method Based on TSP Ant Colony Optimization

Contact Info

Product

Resources

About