Air temperature is one of the main factors for describing the weather behaviour in the earth. Since Indonesia is located on and near equator, then monitoring the air temperature is needed to determine either global climate change occurs or not. Climate change can have an impact on biological growth in various fields. For instance, climate change can affect the quality of production and growth of animal and plants. Therefore, air temperature prediction is important to meteorologists and Indonesian government to provide information in many sectors. Various prediction algorithms have been used to predict temperature and produce different accuracy. In this study, the deep learning method with Long Short-Term Memory (LSTM) model is used to predict air temperature. Here, the results show that LSTM model with one layer and Adaptive Moment Estimation (ADAM) optimizer produce accuracy which is 32% of , 0.068 of MAE and 0.99 of RMSE. Moreover, here, ADAM optimizer is found better than Stochastic Gradient Descent (SGD) optimizer.
Abstract Estimates of government expenditure for the next period are very important in the government, for instance for the Ministry of Finance of the Republic of Indonesia, because this can be taken into consideration in making policies regarding how much money the government should bear and whether there is sufficient availability of funds to finance it. As is the case in the health, education and social fields, modeling technology in machine learning is expected to be applied in the financial sector in government, namely in making modeling for spending predictions. In this study, it is proposed the application of Long Short-Term Memory (LSTM) Model for expenditure predictions. Experiments show that LSTM model using three hidden layers and the appropriate hyperparameters can produce Mean Square Error (MSE) performance of 0.2325, Root Mean Square Error (RMSE) of 0.4820, Mean Average Error (MAE) of 0.3292 and Mean Everage Presentage Error (MAPE) of 0.4214. This is better than conventional modeling using the Auto Regressive Integrated Moving Average (ARIMA) as a comparison model. ; Perkiraan pengeluaran belanja pemerintah untuk periode kedepan merupakan hal yang sangat penting di pemerintah dalam hal ini pada Kementerian Keuangan Republik Indonesia, karena hal ini dapat dijadikan bahan pertimbangan dalam mengambil kebijakan terkait berapa nilai uang yang harus ditanggung pemerintah serta apakah ada ketersediaan dana yang cukup dalam membiayai belanja tersebut untuk periode yang akan datang. Seperti halnya pada bidang kesehatan, pendidikan dan sosial, teknologi pemodelan pada Machine Learning diharapkan dapat diterapkan di bidang keuangan pada pemerintahan, yaitu dalam membuat pemodelan untuk prediksi belanja. Pada penelitian ini, diusulkan penerapan model Long Short-Term Memory (LSTM) untuk prediksi belanja. Eksperimen menunjukkan model LSTM dengan menggunakan tiga hidden layers dan hyperparameter yang tepat dapat menghasilkan performa Mean Square Error (MSE) sebesar 0.2325, Root Mean Square Error (RMSE) sebesar 0.4820, Mean Average Error (MAE) sebesar 0.3292 dan Mean Everage Presentage Error (MAPE) sebesar 0.4214. Ini lebih baik dibandingkan pemodelan konvensional menggunakan Auto Regressive Integrated Moving Average (ARIMA) sebagai model pembanding.
The text classification process has been well studied, but there are still many improvements in the classification and feature preparation, which can optimize the performance of classification for specific applications. In the paper we implemented dictionary based approach and long-short term memory approach. In the first approach, dictionaries will be padded based on field's specific input and use automation technology to expand. The second approach, long short term memory used word2vec technique. This will help us in getting a comprehensive pipeline of end-to-end implementations. This is useful for many applications, such as sorting emails which are spam or ham, classifying news as political or sports-related news, etc
In the marine environment, shore-based radars play an important role in military surveillance and sensing. Sea clutter is one of the main factors affecting the performance of shore-based radar. Affected by marine environmental factors and radar parameters, the fluctuation law of sea clutter amplitude is very complicated. In the process of training a sea clutter amplitude prediction model, the traditional method updates the model parameters according to the current input data and the parameters in the current model, and cannot utilize the historical information of sea clutter amplitude. It is only possible to learn the short-term variation characteristics of the sea clutter. In order to learn the long-term variation law of sea clutter, a sea clutter prediction system based on the long short-term memory neural network is proposed. Based on sea clutter data collected by IPIX radar, UHF-band radar and S-band radar, the experimental results show that the mean square error of this prediction system is smaller than the traditional prediction methods. The sea clutter suppression signal is extracted by comparing the predicted sea clutter data with the original sea clutter data. The results show that the proposed sea clutter prediction system has a good effect on sea clutter suppression.
To protect the environment and achieve the Sustainable Development Goals (SDGs), reducing greenhouse gas emissions has been actively promoted by global governments. Thus, clean energy, such as wind power, has become a very important topic among global governments. However, accurately forecasting wind power output is not a straightforward task. The present study attempts to develop a fuzzy seasonal long short-term memory network (FSLSTM) that includes the fuzzy decomposition method and long short-term memory network (LSTM) to forecast a monthly wind power output dataset. LSTM technology has been successfully applied to forecasting problems, especially time series problems. This study first adopts the fuzzy seasonal index into the fuzzy LSTM model, which effectively extends the traditional LSTM technology. The FSLSTM, LSTM, autoregressive integrated moving average (ARIMA), generalized regression neural network (GRNN), back propagation neural network (BPNN), least square support vector regression (LSSVR), and seasonal autoregressive integrated moving average (SARIMA) models are then used to forecast monthly wind power output datasets in Taiwan. The empirical results indicate that FSLSTM can obtain better performance in terms of forecasting accuracy than the other methods. Therefore, FSLSTM can efficiently provide credible prediction values for Taiwan's wind power output datasets.
PURPOSE: Morse code as a form of communication became widely used for telegraphy, radio and maritime communication, and military operations, and remains popular with ham radio operators. Some skilled users of Morse code are able to comprehend a full sentence as they listen to it, while others must first transcribe the sentence into its written letter sequence. Morse thus provides an interesting opportunity to examine comprehension differences in the context of skilled acoustic perception. Measures of comprehension and short-term memory show a strong correlation across multiple forms of communication. This study tests whether this relationship holds for Morse and investigates its underlying basis. Our analyses examine Morse and speech immediate serial recall, focusing on established markers of echoic storage, phonological-articulatory coding, and lexical-semantic support. We show a relationship between Morse short-term memory and Morse comprehension that is not explained by Morse perceptual fluency. In addition, we find that poorer serial recall for Morse compared to speech is primarily due to poorer item memory for Morse, indicating differences in lexical-semantic support. Interestingly, individual differences in speech item memory are also predictive of individual differences in Morse comprehension. CONCLUSIONS: We point to a psycholinguistic framework to account for these results, concluding that Morse functions like "reading for the ears" (Maier et al., 2004) and that underlying differences in the integration of phonological and lexical-semantic knowledge impact both short-term memory and comprehension. The results provide insight into individual differences in the comprehension of degraded speech and strategies that build comprehension through listening experience. SUPPLEMENTAL MATERIAL: https://doi.org/10.23641/asha.16451868
Air pollution levels have risen as an outcome of urban and industrial development in so many developing countries. People and governments all around the world are concerned about air pollution, which has a severe influence on both personal health and long-term global development. As a government, it is responsible for preventing and controlling air pollution, as well as monitoring the pollutant's impacts on human health. There are numerous computer models available, ranging from statistics to artificial intelligence. Pollution levels are still out of control in some parts of the world due to a wide range of sources and factors. Because of accurate estimates of future air pollution, the government can take necessary action. Forecasting air pollution levels based on environmental data is becoming increasingly relevant as people become more worried about global warming and urban sustainability. For replicating the complicated linkages between these variables, advanced Deep Learning (DL) algorithms hold enormous promise. The objective of this work is to provide a high level of accurate solution to the air pollution forecasting problem. Kaggle data will be employed to train a DL model that will forecast air pollution levels.
The decline and increase in the price of shares of plantation companies is a problem for investors in making decisions to buy or sell shares. Factors influencing the movement of plantation stock prices include CPO commodity price fluctuations, world oil price fluctuations, Rupiah exchange rate fluctuations, government regulations and policies, demands from importing countries, and climate. Forecasting stock prices is expected to help investors to deal with uncertainty in the movement of plantation stock prices. This study applies the Long Short-Term Memory (LSTM) to predict the stock prices of plantation companies using SSMS, LSIP, and SIMP share price data from the period 1 July 2014 - 22 July 2019. Based on the results of the study it was found that the best LSTM model on SSMS shares by using the RMSProp optimizer and 70 hidden neurons produced an RMSE value of 21,328. Then the best LSTM model on LSIP stock by using Adam optimizer and 80 hidden neurons produces an RMSE value of 33,097. Whereas the best LSTM model on SIMP shares using Adamax optimizer and 100 hidden neurons produced an RMSE value of 8,3337.
An earthquake is a tremor felt on the surface of the earth created by the movement of the major pieces of its outer shell. Till now, many attempts have been made to forecast earthquakes, which saw some success, but these attempted models are specific to a region. In this paper, an earthquake occurrence and location prediction model is proposed. After reviewing the literature, long short-term memory (LSTM) is found to be a good option for building the model because of its memory-keeping ability. Using the Keras tuner, the best model was selected from candidate models, which are composed of combinations of various LSTM architectures and dense layers. This selected model used seismic indicators from the earthquake catalog of Bangladesh as features to predict earthquakes of the following month. Attention mechanism was added to the LSTM architecture to improve the model's earthquake occurrence prediction accuracy, which was 74.67%. Additionally, a regression model was built using LSTM and dense layers to predict the earthquake epicenter as a distance from a predefined location, which provided a root mean square error of 1.25. ; Validerad;2021;Nivå 2;2021-04-19 (alebob); Finansiär: Informationand Communication Technology division of the Governmentof the People's Republic of Bangladesh
The implementation of the Covid-19 vaccination carried out by Indonesian government was ignited pros and contras among the public. Certainly, there will be pros and cons about the vaccination from the community. This attituded of pros and cons, which is also called sentiment, can influence people to accept or refuse to be vaccinated. Todays, people express their sentiment in social media in comments, post, or status. One of the methods used to detect sentiment on social media, whether positive or negative, is through a categorisation of text approach. This research provides a deep learning technique for sentiment classification on Twitter that uses Long Short Term Memory (LSTM), for positive, neutral and negative classes. The word2vec word embeddings was used as input, using the pretrained Bahasa Indonesia model from Wikipedia corpus. On the other hand, the topic-based word2vec model was also trained from the Covid-19 vaccination sentiment dataset which collected from Twitter. The data used after balanced is 2564 training data, 778 data validation data, and 400 test data with 1802 neutral data, 1066 negative data, and 566 positive data. The best results from various parameter processes give an F1-Score value of 54% on the test data, with an accuracy of 66%. The result of this research is a model that can classify sentiments with new sentences.
The increasing use of social media and the idea of extracting meaningful expressions from renewable and usable data which is one of the basic principles of data mining has increased the popularity of Sentiment Analysis which is an important working area recently and has expanded its usage areas. Compiled messages shared from social media can be meaningfully labeled with sentiment analysis technique. Sentiment analysis objectively indicates whether the expression in a text is positive, neutral, or negative. Detecting Arabic tweets will help for politicians in estimating universal incident-based popular reports and people's comments. In this paper, classification was conducted on sentiments twitted in the Arabic language. The fact that Arabic has twisted language features enabled it to have a morphologically rich structure. In this paper we have used the Long Short Term Memory (LSTM), a widely used type of the Recurrent Neural Networks (RNNs), to analyze Arabic twitter user comments. Compared to conventional pattern recognition techniques, LSTM has more effective results in terms of having less parameter calculation, shorter working time and higher accuracy.
Solar radiation is among the renewable resources on which modern society relies to partially replace the existing fossil fuel-based energy resources. Awareness of how the energy is produced must complement awareness of how it is consumed. In the economic context, the gains derive from predictability across the entire supply chain. This paper represents a compressive study on how standard recurrent neural networks, long short-term memory, and gated recurrent units can be used to forecast power production of photovoltaic (PV) systems. This approach can be used for other use cases in solar or even wind power prediction since it provides solid fundamentals for working with weather data and recurrent artificial neural networks, being the core of any smart grid management system. Few studies have explored how these models should be implemented, and even fewer have compared the outcomes of different model types. The data used consist of weather and power production data with a one-hour resolution. The data were further pre-processed to unveil the maximum information. The most effective model parameters were selected to make the forecast. Solar energy plays a key role among other renewable energy sources in the European Union's climate action and the European Green Deal. Under these initiatives, important regulations are implemented and financial resources made available for those who possess the capabilities required to solve the open points. The much-needed predictability that gives the flexibility and robustness needed for deploying and adopting more renewable technologies can be ensured by utilizing a neural-based predictive approach.
In the European Union States, household energy usage accounts on average for 40% of overall energy consumption and is responsible for a considerable amount of carbon dioxide emissions. The urgent need to take concrete action to identify solutions that can ensure more effective usage of energy in households, both because of environmental and political reasons, has been repeatedly stated by the European Parliament. White box, grey box and black box predictive models were demonstrated to be a feasible approach to predict the indoor temperature to implement an effective energy management strategy. This study has the purpose of illustrating the potentiality of an LSTM Artificial Neural Network in a short and long-term prediction of the indoor temperature in 15 offices distributed on three storeys of an existing building (Energy Center of Turin (Italy)). The indoor temperature was predicted two hours, five hours and one entire day ahead. The performance of these algorithms has been evaluated not only based on two main criteria (i.e., Root Mean Squared Error and Mean Absolute error) but also by considering the adaptability of the model between the three floors and in terms of different years. Moreover, the proposed work explains how parameters affect performances, aiming to properly identify the optimal model structure. Current results indicate that these models can provide accurate predictions for all the proposed time scales and could all potentially be used for predictive control purposes to optimise the energy demand. The novelty of this study is to show that these models can only be trained on data for a limited period and a specific plane, and then be reliable in predicting indoor temperature, both for different planes and for random periods, taking into account temperature and relative humidity. Furthermore, input parameters are limited to indoor HVAC variables, to ensure acceptable predictions regardless of outdoor parameters availability. The only exception is the outdoor temperature, because of its undeniable ...
Jakarta, the capital region of Indonesia, is experiencing recurring floods, with the most extensive recording loss as high as 350 million dollars. Katulampa Barrage's observation of the Upper Ciliwung River plays a central role in reducing the risk of flooding in Jakarta, especially flowing through the Ciliwung River. The peak flow measured in the barrage would travel 13–14 h to the heart of the city, providing adequate time for the government officials and the residents to prepare for the flood risk. However, Jakarta is continually pressed by the population growth, averaging 1.27% in the past 20 years. The constant growth of Jakarta's population continually develops slums in increasingly inconvenient locations, including the riverbanks, increasing vulnerability to floods. This situation necessitates a more advanced early warning system that could provide a longer forecasting lead time. Satellite remote sensing data propose a promising utility to extend the prediction lead time of extreme events. In the case of this study, Sadewa data is used to predict the water level of Katulampa Barrage using long short-term memory (LSTM) recurrent neural networks (RNN). The results show that the model could predict Katulampa Water Level accurately. The model presents a potential for implementation and additional lead time to increase flood mitigation preparedness.
In order to formulate the long-term and short-term development plans to meet the energy needs, there is a great demand for accurate energy forecasting. Energy autonomy helps to decompose a large-scale grid control into a small sized decisions to attain robustness and scalability through energy independence level of a country. Most of the existing energy demand forecasting models predict the amount of energy at a regional or national scale and failed to forecast the demand for power generation for small-scale decentralized energy systems, like micro grids, buildings, and energy communities. A novel model called Sailfish Whale Optimization-based Deep Long Short- Term memory (SWO-based Deep LSTM) to forecast electricity demand in the distribution systems is proposed. The proposed SWO is designed by integrating the Sailfish Optimizer (SO) with the Whale Optimization Algorithm (WOA). The Hilbert-Schmidt Independence Criterion (HSIC) is applied on the dataset, which is collected from the Central electricity authority, Government of India, for selecting the optimal features using the technical indicators. The proposed algorithm is implemented in MATLAB software package and the study was done using real-time data. The optimal features are trained using Deep LSTM model. The results of the proposed model in terms of install capacity prediction, village electrified prediction, length of R & D lines prediction, hydro, coal, diesel, nuclear prediction, etc. are compared with the existing models. The proposed model achieves percentage improvements of 10%, 9.5%,6%, 4% and 3% in terms of Mean Squared Error (MSE) and 26%, 21%, 16%, 12% and 6% in terms of Root Mean Square Error (RMSE) for Bootstrap-based Extreme Learning Machine approach (BELM), Direct Quantile Regression (DQR), Temporally Local Gaussian Process (TLGP), Deep Echo State Network (Deep ESN) and Deep LSTM respectively. The hybrid approach using the optimization algorithm with the deep learning model leads to faster convergence rate during the training process and enables the small-scale decentralized systems to address the challenges of distributed energy resources. The time series datasets of different utilities are trained using the hybrid model and the temporal dependencies in the sequence of data are predicted with point of interval as 5 years-head. Energy autonomy of the country till the year 2048 is assessed and compared.