Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.

You are viewing the site in preview mode

Skip to main content

Table 4 Comparison of Accuracy between SVR [31], MLP [32], RAQP [34], Vlachogianni (Vlacho) [4], LSTM [35], BLSTM [36], SLSTM [37], and the proposed BlaSt Models

From: Ecosense: a revolution in urban air quality forecasting for smart cities

APCs

Method

1hr

2hr

3hr

4hr

5hr

6hr

7hr

8hr

9hr

10hr

11hr

12hr

\(PM_{2.5}\)

SVR

0.949

0.890

0.859

0.789

0.759

0.731

0.717

0.707

0.699

0.695

0.690

0.680

 

MLP

0.958

0.910

0.890

0.886

0.805

0.778

0.748

0.739

0.723

0.715

0.699

0.690

 

RAQP

0.933

0.925

0.912

0.903

0.880

0.880

0.862

0.775

0.740

0.739

0.725

0.712

 

Vlacho

0.867

0.828

0.792

0.770

0.748

0.719

0.651

0.636

0.621

0.615

0.605

0.590

 

LSTM

0.938

0.928

0.919

0.890

0.860

0.832

0.803

0.791

0.770

0.718

0.712

0.700

 

BLSTM

0.918

0.903

0.899

0.898

0.859

0.835

0.825

0.770

0.739

0.718

0.713

0.705

 

SLSTM

0.969

0.965

0.935

0.912

0.883

0.853

0.833

0.815

0.790

0.758

0.750

0.721

 

BlaSt

0.995

0.985

0.920

0.900

0.900

0.890

0.888

0.868

0.856

0.857

0.835

0.829

\(NO_{2}\)

SVR

0.930

0.925

0.910

0.898

0.893

0.883

0.873

0.867

0.853

0.808

0.797

0.788

 

MLP

0.935

0.870

0.848

0.842

0.809

0.785

0.769

0.759

0.729

0.710

0.700

0.693

 

RAQP

0.966

0.952

0.938

0.910

0.894

0.877

0.862

0.838

0.799

0.788

0.758

0.708

 

Vlacho

0.888

0.859

0.850

0.845

0.833

0.812

0.802

0.795

0.759

0.730

0.723

0.713

 

LSTM

0.962

0.956

0.945

0.915

0.887

0.866

0.852

0.790

0.770

0.761

0.747

0.728

 

BLSTM

0.987

0.964

0.940

0.915

0.894

0.865

0.847

0.810

0.799

0.770

0.756

0.728

 

SLSTM

0.980

0.978

0.960

0.935

0.893

0.846

0.842

0.820

0.799

0.770

0.757

0.740

 

BlaSt

0.995

0.990

0.892

0.898

0.897

0.869

0.860

0.849

0.845

0.839

0.839

0.830

CO

SVR

0.935

0.890

0.889

0.864

0.844

0.818

0.788

0.772

0.760

0.703

0.688

0.680

 

MLP

0.955

0.950

0.940

0.885

0.868

0.838

0.808

0.782

0.765

0.742

0.735

0.723

 

RAQP

0.962

0.943

0.910

0.895

0.883

0.863

0.847

0.829

0.812

0.784

0.760

0.748

 

Vlacho

0.870

0.869

0.865

0.860

0.859

0.847

0.839

0.821

0.813

0.802

0.790

0.782

 

LSTM

0.975

0.944

0.910

0.863

0.853

0.830

0.797

0.760

0.733

0.730

0.727

0.713

 

BLSTM

0.991

0.964

0.940

0.908

0.894

0.866

0.827

0.810

0.790

0.770

0.747

0.718

 

SLSTM

0.992

0.957

0.938

0.910

0.868

0.849

0.827

0.795

0.765

0.735

0.727

0.710

 

BlaSt

0.996

0.968

0.925

0.907

0.899

0.885

0.879

0.869

0.859

0.849

0.839

0.835

  1. Bold values indicate the efficiency as well as the superior performance of our proposed model