All the architectures tested achieved a mean squared error (MSE) oscil translation - All the architectures tested achieved a mean squared error (MSE) oscil Indonesian how to say

All the architectures tested achiev

All the architectures tested achieved a mean squared error (MSE) oscillating close to 0.07.
The increase in the architecture complexity, by adding neurons in the hidden layer, did not
result in the improvement of the test performances. Hence, simple architectures, with just four
neurons in the hidden layer are more appropriated for the solution of the studied problem,
given that smaller networks are faster to train and have lower risk of overfitting (Fiszelew et
al., 2007).
Slicing the output results, and considering values over or equal to 0.5 forest fire regions,
and values lower than 0.5, regions with no fire risk, it was possible to evaluate the test results
accuracy.
The global accuracy ranged around 90%, while the accuracy in the forest samples was
close to 100%. As expected after analyzing the NDVI temporal profile of the samples, the
main errors happened in distinguishing the agricultural areas from the areas that will be
burned. Even so, the accuracy in identifying those areas was still satisfactory, reaching 85%
for the areas with forest fire risk, and 90% for the agricultural areas. As seeing in the MSE
analysis, no improvement in the accuracies occurred with the increase of neurons numbers in
the hidden layer.
Hence, the network architecture chose to perform the simulation in Porto dos Gauchos
municipality was [5 4 1], that is, five variables in the input layer, four neurons in the hidden
layer and one neuron in the output layer.
The results of the simulation in Porto dos Gauchos municipality are displayed in Figure 3.
The initial output values from the ANN model showed several noises, especially in areas that
were not represented in the networks training samples, as boundaries of deforested areas,
rivers and roads (Figure 3a and 3c). To get around this problem, a median filter (kernel size
5x5) was applied to smooth the result values (Jensen, 2005). The product of this procedure is
0/5000
From: -
To: -
Results (Indonesian) 1: [Copy]
Copied!
All the architectures tested achieved a mean squared error (MSE) oscillating close to 0.07. The increase in the architecture complexity, by adding neurons in the hidden layer, did not result in the improvement of the test performances. Hence, simple architectures, with just four neurons in the hidden layer are more appropriated for the solution of the studied problem, given that smaller networks are faster to train and have lower risk of overfitting (Fiszelew et al., 2007).Slicing the output results, and considering values over or equal to 0.5 forest fire regions, and values lower than 0.5, regions with no fire risk, it was possible to evaluate the test results accuracy.The global accuracy ranged around 90%, while the accuracy in the forest samples was close to 100%. As expected after analyzing the NDVI temporal profile of the samples, the main errors happened in distinguishing the agricultural areas from the areas that will be burned. Even so, the accuracy in identifying those areas was still satisfactory, reaching 85% for the areas with forest fire risk, and 90% for the agricultural areas. As seeing in the MSE analysis, no improvement in the accuracies occurred with the increase of neurons numbers in the hidden layer.Hence, the network architecture chose to perform the simulation in Porto dos Gauchos municipality was [5 4 1], that is, five variables in the input layer, four neurons in the hidden layer and one neuron in the output layer. The results of the simulation in Porto dos Gauchos municipality are displayed in Figure 3. The initial output values from the ANN model showed several noises, especially in areas that were not represented in the networks training samples, as boundaries of deforested areas, rivers and roads (Figure 3a and 3c). To get around this problem, a median filter (kernel size 5x5) was applied to smooth the result values (Jensen, 2005). The product of this procedure is
Being translated, please wait..
Results (Indonesian) 2:[Copy]
Copied!
Semua arsitektur diuji mencapai mean kuadrat error (MSE) berosilasi dekat dengan 0,07.
Kenaikan kompleksitas arsitektur, dengan menambahkan neuron pada lapisan tersembunyi, tidak
menghasilkan peningkatan kinerja tes. Oleh karena itu, arsitektur sederhana, dengan hanya empat
neuron pada lapisan tersembunyi lebih disesuaikan untuk solusi dari masalah yang diteliti,
mengingat bahwa jaringan yang lebih kecil lebih cepat untuk melatih dan memiliki risiko lebih rendah overfitting (Fiszelew et
al., 2007).
Mengiris output hasil, dan mempertimbangkan nilai-nilai lebih dari atau sama dengan daerah kebakaran 0,5 hutan,
dan nilai-nilai yang lebih rendah dari 0,5, daerah tanpa risiko kebakaran, adalah mungkin untuk mengevaluasi hasil tes
akurasi.
keakuratan global yang berkisar sekitar 90%, sedangkan akurasi di hutan sampel adalah
mendekati 100%. Seperti yang diharapkan setelah menganalisis NDVI profil temporal sampel, yang
kesalahan utama yang terjadi dalam membedakan daerah pertanian dari daerah yang akan
dibakar. Meski begitu, keakuratan dalam mengidentifikasi daerah-daerah masih memuaskan, mencapai 85%
untuk daerah dengan resiko kebakaran hutan, dan 90% untuk daerah pertanian. Seperti melihat di MSE
analisis, tidak ada perbaikan dalam akurasi terjadi dengan meningkatnya angka neuron di
lapisan tersembunyi.
Oleh karena itu, arsitektur jaringan memilih untuk melakukan simulasi di Porto dos Gauchos
kota itu [5 4 1], yaitu, lima variabel dalam lapisan input, empat neuron di hidden
layer dan satu neuron pada lapisan output.
hasil simulasi di Porto dos Gauchos kotamadya ditampilkan pada Gambar 3.
nilai-nilai keluaran awal dari model ANN menunjukkan beberapa suara, terutama di daerah yang
tidak terwakili dalam sampel pelatihan jaringan, sebagai batas-batas daerah gundul,
sungai dan jalan (Gambar 3a dan 3c). Untuk menyiasati masalah ini, median filter yang (kernel ukuran
5x5) diterapkan untuk kelancaran nilai hasil (Jensen, 2005). Produk dari prosedur ini adalah
Being translated, please wait..
 
Other languages
The translation tool support: Afrikaans, Albanian, Amharic, Arabic, Armenian, Azerbaijani, Basque, Belarusian, Bengali, Bosnian, Bulgarian, Catalan, Cebuano, Chichewa, Chinese, Chinese Traditional, Corsican, Croatian, Czech, Danish, Detect language, Dutch, English, Esperanto, Estonian, Filipino, Finnish, French, Frisian, Galician, Georgian, German, Greek, Gujarati, Haitian Creole, Hausa, Hawaiian, Hebrew, Hindi, Hmong, Hungarian, Icelandic, Igbo, Indonesian, Irish, Italian, Japanese, Javanese, Kannada, Kazakh, Khmer, Kinyarwanda, Klingon, Korean, Kurdish (Kurmanji), Kyrgyz, Lao, Latin, Latvian, Lithuanian, Luxembourgish, Macedonian, Malagasy, Malay, Malayalam, Maltese, Maori, Marathi, Mongolian, Myanmar (Burmese), Nepali, Norwegian, Odia (Oriya), Pashto, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Samoan, Scots Gaelic, Serbian, Sesotho, Shona, Sindhi, Sinhala, Slovak, Slovenian, Somali, Spanish, Sundanese, Swahili, Swedish, Tajik, Tamil, Tatar, Telugu, Thai, Turkish, Turkmen, Ukrainian, Urdu, Uyghur, Uzbek, Vietnamese, Welsh, Xhosa, Yiddish, Yoruba, Zulu, Language translation.

Copyright ©2025 I Love Translation. All reserved.

E-mail: