Grammar-Based Feature Generation for Time-Series Prediction by Anthony Mihirana De Silva, Philip H. W. Leong

By Anthony Mihirana De Silva, Philip H. W. Leong

This publication proposes a unique technique for time-series prediction utilizing computer studying innovations with computerized function iteration. software of desktop studying strategies to foretell time-series maintains to draw huge realization as a result of trouble of the prediction difficulties compounded via the non-linear and non-stationary nature of the true global time-series. The functionality of laptop studying strategies, between different issues, relies on compatible engineering of beneficial properties. This booklet proposes a scientific means for producing appropriate positive factors utilizing context-free grammar. a few characteristic choice standards are investigated and a hybrid characteristic new release and choice set of rules utilizing grammatical evolution is proposed. The e-book comprises graphical illustrations to provide an explanation for the function new release technique. The proposed ways are verified by way of predicting the final rate of significant inventory marketplace indices, height electrical energy load and internet hourly foreign currencies customer exchange quantity. The proposed strategy should be utilized to a variety of desktop studying architectures and purposes to symbolize complicated function dependencies explicitly while computer studying can't accomplish that on its own. business functions can use the proposed strategy to increase their predictions.

Show description

Read Online or Download Grammar-Based Feature Generation for Time-Series Prediction PDF

Similar computer vision & pattern recognition books

Markov Models for Pattern Recognition: From Theory to Applications

Markov types are used to resolve hard trend attractiveness difficulties at the foundation of sequential information as, e. g. , automated speech or handwriting acceptance. This complete advent to the Markov modeling framework describes either the underlying theoretical thoughts of Markov versions - protecting Hidden Markov types and Markov chain types - as used for sequential info and offers the suggestions essential to construct winning platforms for sensible functions.

Cognitive Systems

Layout of cognitive platforms for information to humans poses an important problem to the fields of robotics and synthetic intelligence. The Cognitive structures for Cognitive suggestions (CoSy) undertaking was once equipped to deal with the problems of i) theoretical growth on layout of cognitive platforms ii) equipment for implementation of platforms and iii) empirical reviews to additional comprehend the use and interplay with such platforms.

Motion History Images for Action Recognition and Understanding

Human motion research and popularity is a comparatively mature box, but one that is usually no longer good understood by means of scholars and researchers. the big variety of attainable adaptations in human movement and visual appeal, digicam perspective, and setting, current enormous demanding situations. a few very important and customary difficulties stay unsolved through the pc imaginative and prescient neighborhood.

Data Clustering: Theory, Algorithms, and Applications

Cluster research is an unmanaged approach that divides a suite of items into homogeneous teams. This ebook starts off with uncomplicated info on cluster research, together with the class of knowledge and the corresponding similarity measures, by way of the presentation of over 50 clustering algorithms in teams in keeping with a few particular baseline methodologies akin to hierarchical, center-based, and search-based tools.

Additional resources for Grammar-Based Feature Generation for Time-Series Prediction

Example text

A chromosome is represented as a binary string, with each consecutive group of k-bits creating a codon. A group of codons is called a gene. A chromosome may contain more than one gene. The operations performed on the population are selection, crossover and mutation as illustrated in Fig. 1. The initial population is created by randomly setting each bit of a chromosome. The chromosomes are then evaluated based on a given fitness function φ(·), where each chromosome represents a solution to the formalised problem.

Log(distance) in sqrt(distance^3+log(distance))). To characterise this behaviour, the code was run 100 times and its error from Kepler equation was noted. 40 GHz Intel Core i7-2600 CPU. seed(0) was executed before running the code. 5. Notice that the average performance can be improved at the expense of time, by increasing the GE’s number of generation iterations or population size popSize. GrammaticalEvolution allows monitoring the status of each generation using a callback function. This function, if provided to parameter monitorFunc, receives an object similar to the return value of GrammaticalEvolution.

E. g. branch and bound and beam search. 2. Sequential search: Sequential forward feature selection (SFFS), sequential backward feature elimination (SBFE) and bidirectional selection are greedy search algorithms that add or remove features one at a time [2]. SFFS initiates with an empty set and SBFE initiates with a full set whereas a bidirectional search initiates a SFFS and SBFE simultaneously ensuring that the features selected by SFFS are never eliminated by SBFE. The drawback of SFFS is that it could fail to eliminate redundant features generated in the search process.

Download PDF sample

Rated 4.40 of 5 – based on 48 votes