Research Square (Research Square),
Journal Year:
2024,
Volume and Issue:
unknown
Published: Dec. 6, 2024
Abstract
Machine
learning
tools
were
used
in
this
study
to
extract
information
on
prediction
capabilities
using
regression
and
classification
modalities.
PM10,
PM2.5,
NO,
NO2,
NOX,
NH3,
SO2,
CO,
O3,
Benzene,
Toluene,
Xylene
as
predictors.
AQI
was
a
target
variable
with
numerical
text-encoded
values.
Nineteen
regressor
fifteen
classifier
models
tested
for
capabilities,
features
influencing
presented.
We
six
evaluation
metrics,
i.e.,
MAE,
MSE,
RMSE,
R2,
RMSLE,
MAPE,
under
mode
Accuracy,
AUC,
Recall,
Precision,
F1,
Kappa,
MCC
mode.
When
used,
we
observed
that
the
Extra
Trees
Regressor
performed
well
an
R2
of
0.94.
For
mode,
Random
Forest
Classifier
relatively
better
accuracy
precision
0.824.
PM2.5
PM10
are
vital
essential
conclude
Particulate
matter
is
crucial
predicting
over
stations
considered
supported
by
ML-based
observations.
