Scribd
Upload
45 views

A Study of Soft Computing Techniques

The document summarizes a presentation on using soft computing techniques for time series forecasting. It discusses how time series data is used in various fields like financial forecasting and introduces some common time series forecasting methods like ARIMA. It then focuses on soft computing approaches like artificial neural networks, support vector machines, genetic algorithms, fuzzy logic and how they can be used for time series forecasting and potentially provide more accurate results than single models.

Uploaded by

Pradeep Singla
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
45 views

A Study of Soft Computing Techniques

The document summarizes a presentation on using soft computing techniques for time series forecasting. It discusses how time series data is used in various fields like financial forecasting and introduces some common time series forecasting methods like ARIMA. It then focuses on soft computing approaches like artificial neural networks, support vector machines, genetic algorithms, fuzzy logic and how they can be used for time series forecasting and potentially provide more accurate results than single models.

Uploaded by

Pradeep Singla
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 19

Presentation

On

A Study of soft computing techniques: Case of time


series forecasting

By:
Pardeep Singla
Deenbandhu Chhotu Ram University of Science
& Technology Sonepat
Abstract

• The time series is a series of observed data, recorded at a particular interval


of time.

• This time series are almost used in every field of forecasting such as : financial
forecasting, stock market analysis , wind forecasting , solar forecasting etc.

• The forecasting of near future value have to obtained with lesser errors.

• This paper discusses the various techniques to analyze time series and its
forecasting methods.

• The paper focuses on the various optimization techniques that as to be used


for time series forecast for better accuracy.
Introduction

• Optimization techniques, Specially Soft Computing, are widely used


in numbers of process now days.

• Time series forecasting is one of the important tasks in the area of


power planning, scheduling and balancing.

• In same way, accurate forecasting of time series solar irradiance,


wind power encourages the grid stability.

• In this paper, our goal is to conduct a comparative analysis of all


time series prediction frameworks and how soft computing is used
for dataset.
Time Series Forecasting

• Definition of time series "A time series is a


series of observed data, calculated as a
standard over successive tenures .

• Usually four components are created: Pattern,


Cyclical, Seasonal, and Irregular. These four
factors also represent the statistical pattern
ARIMA- Auto Regression Integrated Moving Average

• ARIMA has become well-liked for regression


forecasting among all.

• In this method, the measurement data presented are


preliminary tested for stationarity and a differentiating
procedure is done if not then.

• On the off chance that the data is still non-stationary,


difference is drawn again before the data is actually
stationary.
• Analyzing the time series
• Plotting of ACF and PACF
• Stationarize time series (If not)
• Finding suitable model specifications
• Parameter estimation: Akaike Information
criteria(AIC) and Bayesian Information
Criteria(BIC)
• Results
Soft Computing Approach Models
Artificial Neural Network (ANN) Model
• ANN model is basically defined by a model of information processing which draws
motivation from the human brain.

• The inspiration of creating ANN method is the creation of a method for demonstration
of the brain that can perform computation.

• ANN can be used in several different activities like: Classification, estimation,


optimization, defining, analysing and managing complex systems .

• A large number of interconnected computing components are used in the neural


network structure (nodes).

• The ANN uses different algorithms like the Levenberg-Marquardt (LM) algorithm, the
scaled conjugate gradient, and the Pola-Ribiere conjugate gradient to predict the
output variables
ANN Structure
Support Vector Machine Model
• It is a type of machine learning that Cortes and Vapnik
introduced with statistical learning in 1995.

• this basic method has been developed for pattern


recognition and is now enthusiastically used for various
technologies such as image retrieval, fault diagnosis,
regression computing and forecasting, etc.

• The time series is used to train the model which is as simple


as that of neural network and there is no problem of curve
over fitting, strucking to local minima in SVM
Genetic algorithms

• It is a computational process with selection process, crossover on the population


and mutation process performed repeatedly on a population set.

• A genetic algorithm (or GA) is a search technique used in computing to find true
or approximate solutions to optimization and search problems.

• This algorithm is not capable of operating on input parameters in its origin form
and can only applied on the coded version of parameters .

• The GA utilizes a series of solutions for population where chromosomes identify


each solution.

• The evolution usually starts from a population of randomly generated individuals


and happens in generations.
• The selection process is first carried out on the basis of the likelihood of
the individual's fitness.

• The new chromosomes were generated from the chosen ones


the stage of reproduction.

• This algorithm utilized mutation and recombination method for the


generation of new chromosomes.

• Then the fitness of the new chromosomes produced is determined.


Finally, the old chromosomes replaced by the new ones when the
population finds an optimal solution, the algorithm are stopped
Flow graph
Fuzzy logic

• Fuzzy logic, introduced by lotfi A. Zadehh in 1965.

• It deals efficiently with unknown datasets.

• It also has the capacity to control the granularity of data and imprecision.

• It follows the function of the cognitive method, on the basis on the principles of
score obtained by relative membership function.

• It takes ambiguous information as input and provides the overall precise decision
as output.

• The value to which ‘0’ represents not belonging and the value to which ‘1’
represents belonging.
Fuzzy Logic Example
Example Rules
IF speed is TOO SLOW and acceleration is DECELERATING,
THEN INCREASE POWER GREATLY

IF speed is SLOW and acceleration is DECREASING,


THEN INCREASE POWER SLIGHTLY

IF distance is CLOSE,
THEN DECREASE POWER SLIGHTLY

...
Time Series Fuzzification

GHIt 1   f (GHI1 ), f (GHI2 ),..., f (GHIt )

f (GHI 1 ) Is the one step before fuzzified time series.


Conclusion
During this article, the time series forecasting and its models
are studied here. The use of soft computing method for time
series prediction such as ANN and SVM was jointly studied.
We learned from the analysis that time series data is a
mixture of linear and nonlinear, so each individual model
does not seem able to handle it properly. However, models
related to the regression such as ARIMA can be use to
forecast the time series but the use of soft computing in
hybridization manner is very popular. Therefore a new time
series analysis space may be a hybridization of models by
leveraging the power of individual models.
References
• [1] Feinberg, E. A., & Genethliou, D. (2005). Load forecasting. Applied mathematics for restructured electric power
systems, 269-285.
• [2] Srivastava, A. K., Pandey, A. S., & Singh, D. (2016). Short-term load forecasting methods: A review. In Emerging
Trends in Electrical Electronics & Sustainable Energy Systems (ICETEESES), International Conference on (pp. 130-
138). IEEE.
• [3] Ali, D., Yohanna, M., Puwu, M. I., & Garkida, B. M. (2016). Long-term load forecast modelling using a fuzzy logic
approach. Pacific Science Review A: Natural Science and Engineering, 18(2), 123-127.
• [4] Adhikari R, Agrawal RK. An Introductory Study on Time Series Modeling and Forecasting, arXiv Prepr
arXiv13026613. 2013.
• [5] Meryem O, Ismail J, Mohammed E-M. A Comparative Study of Predictive Algorithms for Time Series
Forecasting. In: Information Science and Technology (CIST), 2014 Third IEEE International Colloquium; 2014. p.
68−73.
• [6] De Gooijer JG, Hyndman RJ. 25 Years of Time Series Forecasting, Int. J. Forecast. 2006; 22(3):443−73. DOI:
10.1016/j.ijforecast.2006.01.001.
• [7] Babu CN, Reddy BE. A Moving-Average Filter based Hybrid ARIMA–ANN Model for Forecasting Time Series
Data, Appl. Soft. Comput. 2014; 23:27−38. DOI: 10.1016/j.asoc.2014.05.028.
• [8] Hyndman RJ, Athanasopoulos G. Forecasting: Principles and Practice. OTexts; 2014.
• [9] Mandal SN, Choudhury JP, Chaudhuri SRB, De D. Soft Computing Approach in Prediction of a Time Series Data,
J Theor. Appl. Inf. Technol. 2008; 8(3):1131−41.
• [10] Sivanandam, S. N., & Deepa, S. N. (2007). Principles of Soft Computing, John Wiley & Sons.
• [11] Agami N, Atiya A, Saleh M, El-Shishiny H. A Neural Network based Dynamic Forecasting
Model for Trend Impact Analysis, Technol. Forecast Soc. Change. 2009; 76(7):952−62.
• [12] Chen Y, Yang B, Dong J, Abraham A. Time-Series Forecasting using Flexible Neural Tree
Model, Inf. Sci. (Ny). 2005; 174(3):219−35.
• [13] G. H. Hargreaves and Z.A. Samani, “Estimation potential Evapotranspiration,” J.
Irrig.Drain.Div., vol.108, no.3, pp.225-230, 1982.
• [14] Gupta, A., & Sarangi, P. K. (2012). Electrical load forecasting using genetic algorithm based
back-propagation method. ARPN Journal of Engineering and Applied Sciences, 7(8), 1017-1020.
• [15] Heng, E. T. H., Srinivasan, D., & Liew, A. C. (1998, March). Short term load forecasting using
genetic algorithm and neural networks. In Energy Management and Power Delivery, 1998.
Proceedings of EMPD'98. 1998 International Conference on (Vol. 2, pp. 576-581). IEEE.
• [16] Bakirtzis, A. G., Theocharis, J. B., Kiartzis, S. J., & Satsios, K. J. (1995). Short term load
forecasting using fuzzy neural networks. IEEE Transactions on Power Systems, 10(3), 1518-1524.
• [17] Sachdeva, S., & Verma, C. M. (2008, October). Load forecasting using fuzzy methods. In
Power System Technology and IEEE Power India Conference, 2008. POWERCON 2008. Joint
International Conference on (pp. 1-4). IEEE.
Thank You

You might also like