Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

1.

Implement the neuro-fuzzy inference system (ANFIS) model predictor for forecasting of
daily air pollution concentrations of air pollutants (SO2), (NO2), (CO), (O3) (PM10) and
(PM205) based on two datasets. (Note: The dataset will be provided to you after
agreement).
2. Compare the results with standard (please refer to Figure 1). Then, compute the Mean
Absolute Error (MAE), Root Mean Square Error (RMSE), Coefficient of Determination
(R2), computational time
3. GUI to entry and change the data files.

Figure 1: prediction result

Mean Absolute Error (MAE)

According to statistics, Mean Absolute Error (MAE) is a measure of distinction between two
continuous variables. Presume S and H are variables of combined observations that express the
same. Examples of H vs. S consist of contrasts of anticipated vs. observed, subsequent time vs.
first time, as well as one technique of dimension vs. an alternative technique of dimension. MAE
is the average vertical and also horizontal range in between each point as well as the identity line.
The formula for MAE is shown in Equation below:
𝑛
1
𝑀𝐴𝐸 = ∑ |𝑥𝑖−𝑥 | (3.1)
𝑛
𝑖=1

Root-Mean-Square Error (RMSE)


Root Mean Square Error (RMSE) is used to determine the difference between sample values
anticipated by a model. Root Mean Square Error Deviation (RMSD) represents the square root of
the second sample as the difference between the anticipated and observed values or taking the
quadratic mean of these differences. These variances are represented as residuals utilized for
approximating anticipating errors when calculated out of sample. It serves to accumulation the size
of errors numerous times into a single procedure of prediction. It is a measure of accuracy in
comparing forecasting errors in different models of a dataset not in between the datasets as it is a
scale-dependent.

∑𝑛𝑖=1(𝑌𝑖 − 𝑦𝑖 )
𝑅𝑀𝑆𝐸 = √ (3.2)
𝑛

Coefficient of Determination (R2)


The R2 is an important statistical measure which is a regression model that represents the
proportion of the difference or variance in statistical terms for a dependent variable which can be
explained by an independent variable or variables. In short, R2 correlation determines how well
data is fit the regression model or how well the modeled data is fit to observation data. This is the
ability of the prediction model to predict the independent variable. R2 can be calculated by using
Equation below,

𝑛(∑ 𝑥𝑦) − (∑ 𝑥)(∑ 𝑦)


𝑅2 = (3.3)
√[𝑛 ∑ 𝑥 2 − (∑ 𝑥)2 ] [𝑛 ∑ 𝑦 2 − (∑ 𝑦)2 ]

where n is the number in the given dataset x is the first variable in the context (or
observation data) y is the second variable (or modeled data).

Computational Time
Computational time refers to the length of execution time for the prediction process by the model
from the time of receiving the new sample to the time of generating the class prediction of the new
sample.

You might also like