Analysis of canopy phenology in man-made forests using near-earth remote sensing

Abstract Background Forest canopies are highly sensitive to their growth, health, and climate change. The study aims to obtain time sequence images in mix foresters using a near-earth remote sensing method to track the seasonal variation in the color index and select the optimal color index. Three different regions of interest (RIOs) were defined and six color indexes (GRVI, HUE, GGR, RCC, GCC, and GEI) were calculated to analyze the microenvironment difference. The key phenological phase was identified using the double logistic model and the derivative method, and the phenology forecast of color indexes was performed based on the long short-term memory (LSTM) model. Results The results showed that the same color index in different RIOs and different color indexes in the same RIO present a slight difference in the days of growth and the days corresponding to the peak value, exhibiting different phenological phases; the mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) of the LSTM model was 0.0016, 0.0405, 0.0334, and 12.55%, respectively, indicating that this model has a good forecast effect. Conclusions In different areas of the same forest, differences in the micro-ecological environment in the canopies were prevalent, with their internal growth mechanism being affected by different cultivation ways and the external environment. Besides, the optimal color index also varies with species in phenological response, that is, different color indexes are used for different forests. With the data of color indexes as the training set and forecast set, the feasibility of the LSTM model in phenology forecast is verified.

Download Full-text

COVID-19 Infection Forecasting based on Deep Learning in Iran

10.1101/2020.05.16.20104182 ◽

2020 ◽

Cited By ~ 3

Author(s):

Mehdi Azarafza ◽

Mohammad Azarafza ◽

Jafar Tanha

Keyword(s):

Short Term Memory ◽

Mean Squared Error ◽

Moving Average ◽

National Level ◽

Absolute Error ◽

Percentage Error ◽

Autoregressive Integrated Moving Average ◽

Error Metrics ◽

Squared Error ◽

Better Than

Since December 2019 coronavirus disease (COVID-19) is outbreak from China and infected more than 4,666,000 people and caused thousands of deaths. Unfortunately, the infection numbers and deaths are still increasing rapidly which has put the world on the catastrophic abyss edge. Application of artificial intelligence and spatiotemporal distribution techniques can play a key role to infection forecasting in national and province levels in many countries. As methodology, the presented study employs long short-term memory-based deep for time series forecasting, the confirmed cases in both national and province levels, in Iran. The data were collected from February 19, to March 22, 2020 in provincial level and from February 19, to May 13, 2020 in national level by nationally recognised sources. For justification, we use the recurrent neural network, seasonal autoregressive integrated moving average, Holt winter's exponential smoothing, and moving averages approaches. Furthermore, the mean absolute error, mean squared error, and mean absolute percentage error metrics are used as evaluation factors with associate the trend analysis. The results of our experiments show that the LSTM model is performed better than the other methods on the collected COVID-19 dataset in Iran

Download Full-text

RNN-based Dimensional Speech Emotion Recognition

10.31227/osf.io/wa3vp ◽

2020 ◽

Author(s):

Bagus Tris Atmaja

Keyword(s):

Emotion Recognition ◽

Short Term Memory ◽

Mean Squared Error ◽

Absolute Error ◽

Recognition System ◽

Speech Emotion Recognition ◽

Percentage Error ◽

Concordance Correlation ◽

Acoustic Feature ◽

Dense System

◆ A speech emotion recognition system based on recurrent neural networks is developed using long short-term memory networks.◆ Two of acoustic feature sets are evaluated: 31 Features (3 time-domain features, 5 frequency-domain features, 13 MFCCs, 5 F0s, and 5 Harmonics) and eGeMaps feature set (23 features).◆ To evaluate the performance, some metrics are used i.e. mean squared error (MSE), mean absolute percentage error (MAPE), mean absolute error (MAE) and concordance correlation coefficient (CCC). Among those metrics, CCC is main focus as it is used by other researchers.◆ The developed system used multi-task learning to maximize arousal, valence, and dominance at the same time using CCC loss (1 - CCC). The result shows using LSTM networks improve the CCC score compared to baseline dense system. The best CCC score isobtained on arousal followed by dominance and valence.

Download Full-text

Predicting Coordinated Actuated Traffic Signal Change Times using Long Short-Term Memory Neural Networks

Transportation Research Record Journal of the Transportation Research Board ◽

10.1177/03611981211000748 ◽

2021 ◽

pp. 036119812110007

Author(s):

Seifeldeen Eteifa ◽

Hesham A. Rakha ◽

Hoda Eldardiry

Keyword(s):

Loss Function ◽

Short Term Memory ◽

Mean Squared Error ◽

Absolute Error ◽

Loss Functions ◽

Short Term ◽

Mean Relative Error ◽

Squared Error ◽

The Mean ◽

Long Short Term Memory

Vehicle acceleration and deceleration maneuvers at traffic signals result in significant fuel and energy consumption levels. Green light optimal speed advisory systems require reliable estimates of signal switching times to improve vehicle energy/fuel efficiency. Obtaining these estimates is difficult for actuated signals where the length of each green indication changes to accommodate varying traffic conditions and pedestrian requests. This study details a four-step long short-term memory (LSTM) deep learning based methodology that can be used to provide reasonable switching time estimates from green to red and vice versa while being robust to missing data. The four steps are data gathering, data preparation, machine learning model tuning, and model testing and evaluation. The input to the models includes controller logic, signal timing parameters, time of day, traffic state from detectors, vehicle actuation data, and pedestrian actuation data. The methodology is applied and evaluated on data from an intersection in Northern Virginia. A comparative analysis is conducted between different loss functions including the mean squared error, mean absolute error, and mean relative error used in LSTM and a new loss function that is proposed in this paper. The results show that while the proposed loss function outperforms conventional loss functions in overall absolute error values, the choice of the loss function is dependent on the prediction horizon. Specifically, the proposed loss function is slightly outperformed by the mean relative error for very short prediction horizons (less than 20 s) and the mean squared error for very long prediction horizons (greater than 120 s).

Download Full-text

Forecasting Brazilian Ethanol Spot Prices Using LSTM

Energies ◽

10.3390/en14237987 ◽

2021 ◽

Vol 14 (23) ◽

pp. 7987

Author(s):

Gustavo Carvalho Santos ◽

Flavio Barboza ◽

Antônio Cláudio Paschoarelli Veiga ◽

Mateus Ferreira Silva

Keyword(s):

Short Term Memory ◽

Mean Squared Error ◽

Percentage Error ◽

Spot Prices ◽

Statistical Measures ◽

Squared Error ◽

Proposed Model ◽

Price Trends ◽

Long Short Term Memory ◽

Crisis Scenarios

Ethanol is one of the most used fuels in Brazil, which is the second-largest producer of this biofuel in the world. The uncertainty of price direction in the future increases the risk for agents operating in this market and can affect a dependent price chain, such as food and gasoline. This paper uses the architecture of recurrent neural networks—Long short-term memory (LSTM)—to predict Brazilian ethanol spot prices for three horizon-times (12, 6 and 3 months ahead). The proposed model is compared to three benchmark algorithms: Random Forest, SVM Linear and RBF. We evaluate statistical measures such as MSE (Mean Squared Error), MAPE (Mean Absolute Percentage Error), and accuracy to assess the algorithm robustness. Our findings suggest LSTM outperforms the other techniques in regression, considering both MSE and MAPE but SVM Linear is better to identify price trends. Concerning predictions per se, all errors increase during the pandemic period, reinforcing the challenge to identify patterns in crisis scenarios.

Download Full-text

Analysis of Lumber Prices Time Series Using Long Short-Term Memory Artificial Neural Networks

Forests ◽

10.3390/f12040428 ◽

2021 ◽

Vol 12 (4) ◽

pp. 428

Author(s):

Dercilio Junior Verly Lopes ◽

Gabrielly dos Santos Bobadilha ◽

Amanda Peres Vieira Bedette

Keyword(s):

Stock Prices ◽

Short Term Memory ◽

Mean Squared Error ◽

Absolute Error ◽

Short Term ◽

Term Memory ◽

Squared Error ◽

Dense Output ◽

The Usa ◽

Long Short Term Memory

This manuscript confirms the feasibility of using a long short-term memory (LSTM) recurrent neural network (RNN) to forecast lumber stock prices during the great and Coronavirus disease 2019 (COVID-19) pandemic recessions in the USA. The database was composed of 5012 data entries divided into recession periods. We applied a timeseries cross-validation that divided the dataset into an 80:20 training/validation ratio. The network contained five LSTM layers with 50 units each followed by a dense output layer. We evaluated the performance of the network via mean squared error (MSE), root mean squared error (RMSE), and mean absolute error (MAE) for 30, 60, and 120 timesteps and the recession periods. The metrics results indicated that the network was able to capture the trend for both recession periods with a remarkably low degree of error. Timeseries forecasting may help the forest and forest product industries to manage their inventory, transportation costs, and response readiness to critical economic events.

Download Full-text

PREDIKSI DATA TRANSAKSI PENJUALAN TIME SERIES MENGGUNAKAN REGRESI LSTM

Jurnal Nasional Pendidikan Teknik Informatika (JANAPATI) ◽

10.23887/janapati.v9i1.19140 ◽

2020 ◽

Vol 9 (1) ◽

pp. 1

Author(s):

Marie Luthfi Ashari ◽

Mujiono Sadikin

Keyword(s):

Machine Learning ◽

Time Series ◽

Short Term Memory ◽

Mean Squared Error ◽

Percentage Error ◽

Short Term ◽

Term Memory ◽

Root Mean Squared Error ◽

Squared Error ◽

Long Short Term Memory

Sebagai upaya untuk memenangkan persaingan di pasar, perusahaan farmasi harus menghasilkan produk obat – obatan yang berkualitas. Untuk menghasilkan produk yang berkualitas, diperlukan perencanaan produksi yang baik dan efisien. Salah satu dasar perencanaan produksi adalah prediksi penjualan. PT. Metiska Farma telah menerapkan metode prediksi dalam proses produksi, akan tetapi prediksi yang dihasilkan tidak akurat sehingga menyebabkan tidak optimal dalam memenuhi permintaan pasar. Untuk meminimalisir masalah kurang akuratnya proses prediksi tersebut, dalam penelitian yang disajikan pada makalah ini dilakukan uji coba prediksi menggunakan teknik Machine Learning dengan metode Regresi Long Short Term Memory (LSTM). Teknik yang diusulkan diuji coba menggunakan dataset penjualan produk “X” dari PT. Metiska Farma dengan parameter kinerja Root Mean Squared Error (RMSE) dan MAPE (Mean Absolute Percentage Error). Hasil penelitian ini berupa nilai rata – rata evaluasi error dari pemodelan data training dan data testing. Di mana hasil menunjukan bahwa Regresi LSTM memiliki nilai prediksi penjualan dengan evaluasi model melalui RMSE sebesar 286.465.424 untuk data training dan 187.013.430 untuk data testing. Untuk nilai MAPE sebesar 787% dan 309% untuk data training dan data testing secara berurut.

Download Full-text

Multi-step ahead predictive model for blood glucose concentrations of type-1 diabetic patients

Scientific Reports ◽

10.1038/s41598-021-03341-5 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Syed Mohammed Arshad Zaidi ◽

Varun Chandola ◽

Muhanned Ibrahim ◽

Bianca Romanski ◽

Lucy D. Mastrandrea ◽

...

Keyword(s):

Blood Glucose ◽

Diabetes Management ◽

Mean Squared Error ◽

Absolute Error ◽

Target Range ◽

Percentage Error ◽

Diabetic Patients ◽

Average Percentage ◽

Squared Error

AbstractContinuous monitoring of blood glucose (BG) levels is a key aspect of diabetes management. Patients with Type-1 diabetes (T1D) require an effective tool to monitor these levels in order to make appropriate decisions regarding insulin administration and food intake to keep BG levels in target range. Effectively and accurately predicting future BG levels at multi-time steps ahead benefits a patient with diabetes by helping them decrease the risks of extremes in BG including hypo- and hyperglycemia. In this study, we present a novel multi-component deep learning model that predicts the BG levels in a multi-step look ahead fashion. The model is evaluated both quantitatively and qualitatively on actual blood glucose data for 97 patients. For the prediction horizon (PH) of 30 mins, the average values for root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and normalized mean squared error (NRMSE) are $$23.22 \pm 6.39$$ 23.22 ± 6.39 mg/dL, 16.77 ± 4.87 mg/dL, $$12.84 \pm 3.68$$ 12.84 ± 3.68 and $$0.08 \pm 0.01$$ 0.08 ± 0.01 respectively. When Clarke and Parkes error grid analyses were performed comparing predicted BG with actual BG, the results showed average percentage of points in Zone A of $$80.17 \pm 9.20$$ 80.17 ± 9.20 and $$84.81 \pm 6.11,$$ 84.81 ± 6.11 , respectively. We offer this tool as a mechanism to enhance the predictive capabilities of algorithms for patients with T1D.

Download Full-text

FORECASTING HIGH-FREQUENCY FINANCIAL DATA VOLATILITY VIA NONPARAMETRIC ALGORITHMS: EVIDENCE FROM TAIWAN'S FINANCIAL MARKETS

New Mathematics and Natural Computation ◽

10.1142/s1793005706000543 ◽

2006 ◽

Vol 02 (03) ◽

pp. 345-359 ◽

Cited By ~ 1

Author(s):

WO-CHIANG LEE

Keyword(s):

Computational Intelligence ◽

High Frequency ◽

Mean Squared Error ◽

Absolute Error ◽

Financial Data ◽

Percentage Error ◽

Integrated Volatility ◽

Squared Error ◽

Performance Indexes ◽

Forecasting Performance

This paper uses two computational intelligence algorithms, namely, artificial neural networks (ANN) and genetic programming (GP), for forecasting the volatility of high-frequency TAIEX financial data with four different horizons and compares the out-sample forecasting performance with the GARCH(1,1), EGRACH(1,1) and GJR-GARCH(1,1) models. Based on intraday integrated volatility, the mean squared error (MSE), mean absolute error (MAE), mean absolute percentage error (MAPE), Theil's U and the VaR backtest are used as performance indexes. Our empirical results reveal that the GP and ANN perform reasonably well in forecasting out-sample volatility compared to other parametric volatility forecasting models for most of the performance indexes. Our results also suggest that nonparametric computational intelligence algorithms are powerful for modeling the volatility of high-frequency intraday financial data.

Download Full-text

Radial basis function neural network for 2 satisfiability programming

Indonesian Journal of Electrical Engineering and Computer Science ◽

10.11591/ijeecs.v18.i1.pp459-469 ◽

2020 ◽

Vol 18 (1) ◽

pp. 459 ◽

Cited By ~ 3

Author(s):

Shehab Alzaeemi ◽

Mohd. Asyraf Mansor ◽

Mohd Shareduwan Mohd Kasihmuddin ◽

Saratha Sathasivam ◽

Mustafa Mamat

Keyword(s):

Neural Network ◽

Radial Basis Function ◽

Basis Function ◽

Mean Squared Error ◽

Absolute Error ◽

Percentage Error ◽

Squared Error ◽

Training Technique ◽

Radial Basis ◽

Hidden Neurons

<span>Radial Basis Function Neural Network (RBFNN) is very prominent in data processing. However, improving this technique is vital for the NN training process. This paper presents an integrated 2 Satisfiability in radial basis function neural network (RBFNN-2SAT). There are two different types of training in RBFNN, namely no-training technique and half-training technique. The performance of the solutions via Genetic Algorithm (GA) training was investigated by comparing the Radial Basis Function Neural Network No-Training Technique (RBFNN- 2SATNT) and Radial Basis Function Neural Network Half-Training Technique (RBFNN- 2SATHT). The comparison of both techniques was examined on 2 Satisfiability problem by using a C# software that was developed for this experiment. The performance of the RBFNN-2SATNT and RBFNN-2SATHT in performing 2SAT is discussed in terms of root mean squared error (RMSE), sum squared error (SSE), mean absolute percentage error (MAPE), mean absolute error (MAE), number of the hidden neurons and CPU time. Results obtained from a computer simulation showed that RBFNN-2SATHT outperformed RBFNN-2SATNT.</span>

Download Full-text

Featureless Blood Pressure Estimation Based on Photoplethysmography Signal Using CNN and BiLSTM for IoT Devices

Wireless Communications and Mobile Computing ◽

10.1155/2021/9085100 ◽

2021 ◽

Vol 2021 ◽

pp. 1-10

Author(s):

Yung-Hui Li ◽

Latifa Nabila Harfiya ◽

Ching-Chun Chang

Keyword(s):

Blood Pressure ◽

Short Term Memory ◽

Mean Squared Error ◽

Absolute Error ◽

Feature Engineering ◽

Short Term ◽

Squared Error ◽

Long Short Term Memory ◽

Iot Devices ◽

Blood Pressure Estimation

Continuous blood pressure (BP) acquisition is critical to health monitoring of an individual. Photoplethysmography (PPG) is one of the most popular technologies in the last decade used for measuring blood pressure noninvasively. Several approaches have been carried out in various ways to utilize features extracted from PPG. In this study, we develop a continuous systolic and diastolic blood pressure (SBP and DBP) estimation mechanism without the need for any feature engineering. The raw PPG signal only got preprocessed before being fed to our model which mainly consists of one-dimensional convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) network. We evaluate the resulting SBP and DBP value by the root-mean-squared error (RMSE) and mean absolute error (MAE). This study addresses the effectiveness of the model by outperforming the previous feature engineering-based methods. We achieve RMSE of 11.503 and 6.525 for SBP and DBP, respectively, and MAE of 7.849 and 4.418 for SBP and DBP, respectively. The proposed method is expected to substantially enhance the current efficiency of healthcare IoT (Internet of Things) devices in BP monitoring using PPG signals only.

Download Full-text