lfl (version 1.2)

lfl-package: Linguistic Fuzzy Logic

Description

This package provides functions to work with various algorithms related to linguistic fuzzy logic. Namely, the following functions are available: mining of linguistic fuzzy association rules, performing perception-based logical deduction (PbLD), and forecasting time-series using fuzzy rule-based ensemble (FRBE).

Details

ll{ Package: lfl Type: Package Version: 1.0 Date: 2015-01-01 License: GNU/GPL version 3.0 or later }

This packages provides functions related with linguistic fuzzy logic. To convert data into membership degrees of fuzzy sets that model linguistic expressions, see lcut. To create fuzzy set partitions see fcut.

For fuzzy association rules mining, there is a function searchrules that searches for association rules and computes various statistics about them. There is also reduce to make small rule bases from mined rules by dropping rules that do not contribute singnificantly to the rule base coverage of source data.

To perform Perception-based Logical Deduction (PbLD), please use pbld function. Another ad-hoc inferences may be programmed with help of fire, perceive, aggregate, or defuzz functions.

Fuzzy Rule-based Ensemble (FRBE) is a tool for time-series prediction. Several existing time-series forecasting methods are combined based on features of given time-series to provide a robust forecast -- see frbe. Also evalfrbe may be used to evaluate the performance of forecasting.

Examples

# --- SEARCHING FOR RULES ---
# split data into training and testing set
testing <- CO2[1:5, ]
training <- CO2[-1 * 1:5, ]

# custom context of the RHS variable
uptakeContext <- c(7, 28.3, 46)

# convert training data into fuzzy sets
d <- lcut3(training, context=list(uptake=uptakeContext))

# search for rules
r <- searchrules(d, lhs=1:38, rhs=39:58)



# --- PBLD INFERENCE WITH FOUND RULES ---
# convert testing data info fuzzy sets
x <- lcut3(testing, context=list(uptake=uptakeContext))

# prepare values and partition
v <- slices(uptakeContext[1], uptakeContext[3], 1000)
p <- lcut3(v, name='uptake', context=uptakeContext)

# do the inference
pbld(x, r, p, v)



# --- FRBE TIME-SERIES FORECASTING ---
# prepare data (from the forecast package)
library(forecast)
horizon <- 10
train <- wineind[-1 * (length(wineind)-horizon+1):length(wineind)]
test <- wineind[(length(wineind)-horizon+1):length(wineind)]

# compute forecast
f <- frbe(ts(train, frequency=frequency(wineind)), h=horizon)

# display the forecast
f$mean

# evaluate the results
evalfrbe(f, test)