Back Propagation Neural Network Method Research Articles

Detection of Marine Oil Spills Based on HOG Feature and SVM Classifier

Oil spill accidents have gradually increased due to the continuous development of marine transportation and petroleum processing industries. Monitoring and managing marine oil spills present important economic, social, and practical implications in preventing offshore oil pollution and maintaining ecological balance. Unmanned aerial vehicle (UAV) has become a suitable carrier for low-altitude oil spill detection because of their fast deployment and low cost. Thermal infrared remote sensing images are used as the research object in this study. A method around histogram of gradient (HOG) features combined with a support vector machine (SVM) is proposed for identifying oil spills at sea to improve the accuracy of offshore low-altitude oil spill recognition and realize all-weather monitoring of offshore oil spills in offshore waters. Steps for extracting HOG features and basic principles of the SVM classification are first investigated. Image preprocessing is then performed on collected thermal infrared image data to produce samples. HOG features of samples are extracted, and the radial basis function is selected as the kernel function for training the SVM classifier. HOG features of the infrared image to be tested are calculated and then sent to the classifier for identifying the oil spills. In addition, the proposed method is compared with the back propagation(BP) neural network method and local binary pattern (LBP) combined with the SVM classification method for analysis. The results show that the oil film recognition method based on the HOG feature and SVM has a recognition accuracy of 91.3% in the environment of small infrared oil film samples, which is significantly better than the BP and LBP-SVM recognition methods, and obtains a shorter training time. The method proposed in this study has obvious advantages in terms of small sample size and processing efficiency, can meet the requirements of all-weather inspection of oil film pollutants by UAV in offshore port areas, and has great application potential in the field of maritime supervision informatization in the future.

Optimizing the quality control of multivariate processes under an improved Mahalanobis–Taguchi system

Quality characteristics in manufacturing are correlated and do not follow a normal distribution. This study proposes a quality control method for multivariate manufacturing processes that are based on an improved Mahalanobis–Taguchi System (IMTS). The MTS has no data distribution assumptions and identifies anomalies through the Mahalanobis distance (MD). However, a covariance distance can consider the correlation between variables. Further, to address the shortcomings of the MTS in feature selection and threshold determination. A joint optimization model is proposed in this paper. Under this approach, the IMTS is employed to perform composite analyses on multiple quality characteristics and reduce dimensionality to identify abnormalities and the key quality characteristics that lead to anomalies. Further, various models are compared to construct the optimal non-parametric prediction models for each key quality characteristic. Finally, a conceptual model of process parameter optimization is proposed, which improves the Taguchi method to obtain the optimal combination of process parameters and their importance ranking, as the basis for process adjustment. By applying the proposed method, results show that the IMTS has an abnormality identification rate of 99.5%, which is higher than other methods such as MTS, support vector machine (SVM), back propagation neural network (BPNN), fast correlation-based filter solution SVM (FCBF-SVM) and sequential backward selection BPNN (SBS-BPNN). The dimensionality reduction rate is 0.5, which is higher than MTS, SVM, BPNN, and SBS-BPNN methods. The random forest (RF) algorithm is used for accurate predictions of all five key quality characteristics, the improved Taguchi method guided adjustments to manufacturing processes objectively, effectively, and economically.

Neural Network Back-Propagation Method as Forecasting Technique

Uncontrolled population increase has a direct impact on the development of human resource quality, and the rate of economic growth is becoming increasingly challenging. It was the purpose of the study to apply artificial intelligence tools to examine the forecasts of active family planning participants. Between 2019 and 2020, Pematangsiantar’s family planning participants will be counted as part of the dataset. The dataset was gathered through observation and interviews. The problem was solved using an artificial intelligence technique with a backpropagation algorithm. The analysis is carried out with the aid of Matlab R2011a. Six other architectural models, including 6-4-1, 6-4-2-1, 6-6-8-1, and 6-8-10-1, were evaluated for their suitability. The 6-4-2-1 architectural model on Backpropagation may be used to forecast the number of active family planning participants with a training accuracy rate of 75% and a test accuracy rate of 88%, according to the study’s findings. Results are expected to shed light on the necessity and availability of contraception for Pematangsiantar’s active family planning participants.

Evaluation and Analysis on the Temperature Prediction Model for Bailing Mushroom in Jizhou, Tianjin

Based on the air temperature, wind speed, humidity, air pressure, etc., of the regional automatic weather station in Chutouling Town, Jizhou from April 2019 to November 2020, and the air temperature of the microclimate observation station receiving data every 10 min in a bailing mushroom greenhouse, this paper analyzed and evaluated a BP (back propagation) neural network and stepwise regression method to establish a prediction model for the temperature in the Bailing mushroom greenhouse for different seasons. The results showed that: (1) The air temperature, wind speed, humidity and air pressure outside the shed were the main factors for building the temperature prediction model for the inside temperature, and the air temperature was the most important factor affecting the temperature inside the shed. After introducing humidity, wind speed and air pressure, the accuracy of the model was significantly improved. (2) The temperature prediction model based on the BP neural network method, for every 10 min interval in the greenhouse, for the Bailing mushroom in different seasons, was more accurate than the stepwise regression model. The simulation results of the two models had the highest accuracy in summer, followed by autumn. (3) The root means square error of the BP neural network and stepwise regression model for inside the greenhouse, simulating the daily temperature variations for different seasons, was 1.25, 1.10, 1.08, 1.31 °C and 1.29, 1.19, 1.11, 1.37 °C, respectively. The BP neural network method performed better for predicting the daily temperature variations in seasons. (4) The specifying data of high temperature (24 July 2020) and strong cold wave (31 December 2019) were selected to test the two model methods; the results showed that the simulation of the BP neural network model was better than the stepwise regression model.

A neural network approach for short-term water demand forecasting based on a sparse autoencoder

Abstract This paper presents a backpropagation neural network (BPNN) approach based on the sparse autoencoder (SAE) for short-term water demand forecasting. In this method, the SAE is used as a feature learning method to extract useful information from hourly water demand data in an unsupervised manner. After that, the extracted information is employed to optimize the initial weights and thresholds of the BPNN. In addition, to enhance the effectiveness of the proposed method, data reconstruction is implemented to create suitable samples for the BPNN, and the early stopping method is employed to overcome the BPNN overfitting problem. Data collected from a real-world water distribution system are used to verify the effectiveness of the proposed method, and a comparison with the BPNN and other BPNN-based methods which integrate the BPNN with particle swarm optimization (PSO) and the mind evolutionary algorithm (MEA), respectively, is conducted. The results show that the proposed method can achieve fairly accurate and stable forecasts with a 2.31% mean absolute percentage error (MAPE) and 320 m3/h root mean squared error (RMSE). Compared with the BPNN, PSO–BPNN and MEA–BPNN models, the proposed method gains MAPE improvements of 5.80, 3.33 and 3.89%, respectively. In terms of the RMSE, promising improvements (i.e., 5.27, 2.73 and 3.33%, respectively) can be obtained.

Spectral Reflectance Reconstruction from Red-Green-Blue (RGB) Images for Chlorophyll Content Detection.

Chlorophyll is one of the most important pigments in plants, and the measurement of chlorophyll levels enables real-time monitoring of plant growth, which is of great importance to the vegetation monitoring. Compared with the high cost and time-consuming operation of hyperspectral imaging technique, the spectral reflectance reconstruction technique based on RGB images has the advantages of being inexpensive and fast. In this study, using the example of ginkgo leaves, the spectra were reconstructed from red-green-blue (RGB) images taken by smartphones based on a back propagation (BP) neural network and pseudo-inverse method. Based on a BP neural network, the maximum absolute error between the reconstructed spectra and the reference spectra acquired by the hyperspectral camera was less than 0.038. A partial least squares regression (PLSR) prediction model for chlorophyll content estimation was established using the reconstructed spectra. The R2 and root mean square error (RMSE) of the validation set were 0.8237 and 1.1895%, respectively, there was a high correlation between predicted and measured values. Compared with the pseudo-inverse method, the maximum absolute error of the reconstructed spectra was reduced by 10.9%, the R2 in the chlorophyll prediction results was improved by 12.7%, and the RMSE was reduced by 19.3%. This research showed that reconstructing spectral reflectance based on RGB images can realize real-time measurement of chlorophyll content. It provided a reliable tool for fast and low-cost monitoring of plant physiology and growth conditions.

Bathymetric Prediction Using Multisource Gravity Data Derived From a Parallel Linked BP Neural Network

AbstractGravity anomalies (GAs) and vertical gravity gradient anomalies (VGGs) are regularly used to predict bathymetry. However, few studies have explored the combination of the GAs and VGGs to predict bathymetry. We introduce the back propagation (BP) neural network into bathymetric prediction field and propose a method to predict depth from GAs and VGGs. The method was tested in the Mariana Trench region, and a neural network bathymetry model was constructed using feature data obtained from GAs and VGGs as the input of the BP neural network. Additionally, single‐beam sounding data were used as label data. By comparing the neural network and the gravity‐geologic method (GGM), the neural network was found to provide better performance with an accuracy improvement of 19%. The root‐mean‐square of the absolute difference between the neural network bathymetry model and the single‐beam sounding data was 72.40 m, with a relative accuracy of 1.71%. Approximately 50% of the differences were distributed within ±20 m, and 90% were distributed within ±100 m. The neural network bathymetry model was also compared with the GGM bathymetry model for different depths and topographies, and the results verified the feasibility and effectiveness of the BP neural network method.

The Estimation of Centrifugal Pump Flow Rate Based on the Power–Speed Curve Interpolation Method

During the global energy crisis, it is essential to improve the energy efficiency of pumps by adjusting the pump’s control strategy according to the operational states. However, monitoring the pump’s operational states with the help of external sensors brings both additional costs and risks of failure. This study proposed an interpolation method based on PN curves (power–speed curves) containing information regarding motor shaft power, speed, and flow rate to achieve high accuracy in predicting the pump’s flow rates without flow sensors. The impact factors on the accuracy of the estimation method were analyzed. Measurements were performed to validate the feasibility and robustness of the PN curve interpolation method and compared with the QP and back-propagation neural network (BPNN) methods. The results indicated that the PN curve interpolation method has lower errors than the other two prediction models. Moreover, the average absolute errors of the PN curve interpolation method in the project applications at 47.5 Hz, 42.5 Hz, 37.5 Hz, and 32.5 Hz are 0.1442 m3/h, 0.2047 m3/h, 0.2197 m3/h, and 0.1979 m3/h. Additionally, the average relative errors are 2.0816%, 3.2875%, 3.6981%, and 2.9419%. Hence, this method fully meets the needs of centrifugal pump monitoring and control.

Clear sky color modeling based on BP neural network

Daylight is a valuable resource that provides energy and luminance, as it contains changing spectrum distributions, which also carry color information and have an important impact on human vision and perception. Several sky models have been proposed for use in daylighting simulation calculations to quantify the effect of daylight. However, these models do not fully consider the color, or are modeled using the correlation of the color temperature and luminance fitting, which has color deviation, so it is difficult to realize the chromaticity-related research of daylight. Therefore, the clear sky spectral data in the Beijing area was measured and used to calculate the color coordinates of sky patches. Corresponding it with the sun position, sky patch position, and radiation data to model the sky color distribution based on backpropagation neural network method. Four-, five-, and six-variable models were established, where the five-variable model exhibited the smallest error: NMBEx = 0.52%, CVRMSEx = 3.54%, NMBEy = 0.67%, and CVRMSEy = 3.55%. However, the four- and six-variable errors were also within the specification and were easier to use. The proposed models can achieve the real-time prediction of the sky color distribution, provide a theoretical basis for daylighting design in future buildings, and guide the development of new spectrally-selective material applications and intelligent lighting control systems.

Evaluation of Coalbed Methane Content by Using Kernel Extreme Learning Machine and Geophysical Logging Data

Accurate evaluation of coalbed methane (CBM) content plays a momentous role in the identification and efficient development of favorable exploitation blocks of CBM resources, but there are still many technical challenges in the exploration and development of onshore CBM fields. With the development and application of geophysical logging technology, using geophysical logging data to predict the gas content of CBM reservoirs has been proven to be an effective and feasible solution. However, the complex logging response of the CBM reservoirs makes it difficult to characterize the relationship between the gas content and the logging curve response by a simple linear relationship. In this paper, kernel extreme learning machine (KELM), a machine learning method, is combined with the geophysical logging data to predict the vertical variation curve of gas content in CBM wells. In this paper, the laboratory data on coal rock gas content from 12 CBM wells in the Southern Shizhuang block are selected, and a CBM content prediction model based on the KELM method is constructed by selecting the log curves, combining cross-validation and grid-seeking to determine the hyperparameters, and validating the prediction model using the test dataset and a new well in the same block. The application of the model on the test dataset was remarkable, and the vertical variation of CBM content obtained by applying it to the new well was consistent with the laboratory results, which proved the correctness and generalizability of the model. The results of this paper show that the CBM content evaluation model based on the KELM method and geophysical logging data is applicable to the 3# coal seam in the target block and can be used to predict the vertical CBM content of CBM wells; compared with the extreme learning machine (ELM) method and the backpropagation neural network (BPNN) method, the KELM method requires fewer hyperparameters to be explored when constructing the CBM content evaluation model, and the model construction is simple and has high prediction accuracy. At the same time, the CBM content model constructed by the KELM method differs for different blocks, coal seams at different depths, and different response ranges of geophysical logging data. The construction of a CBM content prediction model using the KELM method and logging curves is an effective means of characterizing CBM resources, and the model construction process and evaluation criteria studied in this paper can be used to help other blocks evaluate the CBM content, providing guidance for further exploration and development of CBM fields with practical application.

Regional Economic Forecasting Method Based on Recurrent Neural Network

Macroeconomic situation is the overall performance of the economic situation of a country and region. Making accurate forecasts of macroeconomic trends is of great significance for analyzing the success or failure of macroeconomic control policies, evaluating the quality of economic system operation, and correctly formulating future development planning strategies. The macroeconomic system is a nonlinear system, the environment is constantly changing, and additional disturbing factors directly affect the operation of the macroeconomic system, which has a great impact on the forecast results. The historical information required for macroeconomic modeling is unstable, unclear, and incomplete, which makes it very difficult to solve such problems with traditional forecasting methods. In response to the multivariate and nonlinear characteristics of macroeconomic forecasting, this paper proposes the application of artificial neural networks for forecasting. This paper introduces the recurrent neural network into the field of economic forecasting to solve the problems of the traditional BP (back propagation) neural network method. The experimental data are verified and the experimental results prove that the studied scheme based PSO-GRU improve the performance of economic forecasting.

An improved Kansei engineering method based on the mining of online product reviews

The explosive growth of online review data on e-commerce websites demands major changes to product emotional design. Kansei engineering, the primary approach of product emotional design, faces multiple problems. This study aims to improve Kansei engineering method to adapt to the emotional design trends of the big data era. Targeting the data of online product reviews, the authors mined text data, calculated fine-grained emotions, and presented an innovative design flow for Kansei engineering. Firstly, a crawler program was compiled to mine online product reviews and product parameters from e-commerce websites, and the mined data were processed. Next, a Kansei word pair screening method was proposed based on term frequency – evaluation, potency, activity (TF-EPA), and used to obtain the Kansei word pairs depicting the imagery of product emotions. After that, a fine-grained calculation method for Kansei evaluation value of products was proposed based on term clustering and adverbs of degrees, and utilized to compute the evaluation value of each product sample under each Kansei word pair. Finally, the backpropagation (BP) neural network was extended into a mapping model between product parameters and Kansei evaluation values, and the prediction performance of the model was evaluated. The smart phone was taken as an example to verify the proposed method, The optimal training values of the 6 groups of Kansei word pairs were 0.00061, 0.00053, 0.00065, 0.00038, 0.00065, and 0.00093, respectively. The absolute error mean of 28 test sets in the experiment under 6 groups of Kansei word pairs were 0.0696, 0.1001, 0.0784, 0.0815, 0.1328, and 0.1358, respectively. Taking the third group as an example, the absolute error mean obtained by the multiple linear regression method was 0.1127 greater than 0.0784, which proved that the BP neural network method used in this paper was superior to the multiple linear regression method in accuracy.

Forecasting Crop Residue Fires in Northeastern China Using Machine Learning

With repeated changes to local crop residue disposal policies in recent years, the distribution and density of crop residue fire events have been irregular in both space and time. A nonlinear and complex relationship between natural and anthropogenic factors often affects the occurrence of crop residue field fires. To overcome this difficulty, we used the Himawari-8 wildfire data for 2018–2021 to verify the likelihood of crop residue fires against the results of three machine learning methods: logistic regression, backpropagation neural network (BPNN), and decision tree (DT). The results showed the verified accuracies of BPNN and DT methods were 68.59 and 79.59%. Meantime, the sensitivity and specificity of DT performed the best, with the value of area under the curve (AUC) 0.82. Furthermore, among all the influencing factors, open burning prohibition constraints, relative humidity and air pressure showed significant correlations with open burning events. As such, BPNN and DT could accurately forecast the occurrence of agricultural fires. The results presented here may improve the ability to forecast agricultural field fires and provide important advances in understanding fire formation in Northeastern China. They would also provide scientific and technical support for crop fire control and air quality forecasting.

Combination of Feature Extractions for Classification of Coral Reef Fish Types Using Backpropagation Neural Network

Feature extraction is important to obtain information in digital images, where feature extraction results are used in the classification process. The success of a study to classify digital images is highly dependent on the selection of the feature extraction method used, from several studies providing a combination of feature extraction solutions to produce a more accurate classification. Classifying the types of marine fish is done by identifying fish based on special characteristics, and it can be through a description of the shape, fish body pattern, color, or other characteristics. This study aimed to classify coral reef fish species based on the characteristics contained in fish images using Backpropagation Neural Network (BPNN) method. Data used in this research was collected directly from Bunaken National Marine Park (BNMP) in Indonesia. The first stage was to extract shape features using the Geometric Invariant Moment (GIM) method, texture features using Gray Level Co-occurrence Matrix (GLCM) method, and color feature extraction using Hue Saturation Value (HSV) method. The third value of feature extraction was used as input for the next stage, namely the classification process using the BPNN method. The test results using 5-fold cross-validation found that the lowest test accuracy was 85%, the highest was 100%, and the average was 96%. This means that the intelligent model derived from the combination of the three feature extraction methods implemented in the BPNN training algorithm is very good for classifying coral reef fish.

Engine RPM and Battery SOC Activation Optimization in Hybrid Vehicle Energy Management System Utilizing BPNN - Genetic Algorithm and BPNN – Particle Swarm Optimization

The energy used in the hybrid vehicle needs to be regulated to gain further mileage and lower fuel consumption. It is achieved by selecting the correct levels of hybrid energy management system (EMS) parameters (i.e., vehicle speed, engine RPM, and activation State of Charge (SOC) of battery). This study focused on the modeling and optimization of Sepuluh Nopember Institute of Technology (ITS)’s series plug-in hybrid electric vehicle (PHEV) car mileage and fuel consumption by comparing the backpropagation neural network (BPNN) method – genetic algorithm (GA) and BPNN – particle swarm optimization (PSO). The BPNN was used to model the character of ITS’s series PHEV EMS and predict mileage and fuel consumption. The BPNN’s model obtained the best EMS parameters, most extended mileage, and minimum fuel consumption. The result of the validation experiment showed that both the integration of BPNN - GA and BPNN - PSO were able to predict and optimize the multi-objective characteristic with the same results.

Reliability evaluation of corroded pipeline under combined loadings based on back propagation neural network method

Corrosion is a common type of defect that causes failure of oil and gas pipelines. Pipelines often bear the combined loadings of axial compression force, bending moment and internal pressure in operation; therefore, it is of great significance to study the reliability of pipeline under different combined loadings. First, this work established a nonlinear finite element model of corroded pipeline for the typical internal corrosion defects of groove shape, and the reliability of the model and the accuracy of the finite element analysis method are verified. Then the burst pressure of pipeline under combined loadings is investigated. Second, this work established the BP network analysis and prediction model. The results of numerical simulation are used as samples to train the neural network. The prediction limit function value is output through the forward simulation function, which is compared with the objective function to judge whether the pipeline is reliable. Finally, back propagation neural network model calculated the reliability of pipeline under different working conditions combined with Monte Carlo method. The effects of load, corrosion depth, axial corrosion length, circumferential corrosion width and variation coefficient of parameters on pipeline reliability are analyzed. It is found that axial force and bending moment affect the reliability of corroded pipeline. Generally, the higher the combined loadings are, the lower the reliability is, but this law will be different due to the influence of corrosion parameters. In addition, the corrosion depth has the greatest influence on the reliability of corroded pipeline, followed by the corrosion width and the corrosion length. The coefficient of variation of corrosion depth has a great influence on the reliability of pipeline with high depth corrosion defect. These conclusions can provide references for the reliability evaluation of pipelines.

A prediction model of wall shear stress for ultra-high-pressure water-jet nozzle based on hybrid BP neural network

ABSTRACT Two hybrid back-propagation neural network (BPNN) models optimized by two heuristic search algorithms, namely genetic algorithm (GA-BP) and particle swarm optimization (PSO-BP), are proposed in this paper to predict radial maximum wall shear stress instead of traditional computational fluid dynamics (CFD) methods. The two proposed models are trained and validated using a database of 150 radial maximum wall-shear-stress values obtained via CFD simulations. The best fit model is identified from various BPNN models based on the Akaike information criterion (AIC) and the Bayesian information criterion (BIC), balancing the trade-off between goodness-of-fit and model complexity. The model performance is evaluated by MAE, RMSE, regression coefficient (R), and Nash-Sutcliffe efficiency (NSE). The best fit model is a three-layered BPNN model consisting of a 4:4:1 topology. In almost all evaluation indicators, the two hybrid BPNN methods outperform three existing algorithms, namely classical BPNN, random forest (RF), and gradient boosting decision tree (GBDT). Both PSO-BP and GA-BP can provide a more precise assessment of radial maximum wall shear stress, with maximum errors being 5.81% and 8.24% respectively. The proposed PSO-BP prediction model is promising and has great feasibility in predicting the radial maximum wall shear stress of UHP water-jet nozzle in engineering applications.

Regression prediction of hydrogen enriched compressed natural gas (HCNG) engine performance based on improved particle swarm optimization back propagation neural network method (IMPSO-BPNN)

Artificial neural network (ANN) methods have been rapidly developed and applied in solving nonlinear small sample problems. In this paper, an improved particle swarm algorithm optimized back propagation neural network (IMPSO-BPNN) method was proposed and used for the regression analysis and prediction of 20% (volume fraction) hydrogen enriched compressed natural gas (HCNG) engine performance. Meanwhile, various ANN and support vector machine (SVM) methods were also utilized for a comparative study. The experimental results show that the HCNG engine has the highest combustion efficiency, the maximum output torque and the minimum brake specific fuel consumption (BSFC) when operating at the maximum brake torque (MBT) timing, and the brake specific NOx (BSNOx) is also at a relatively low level. Through the comparison of multiple methods, the prediction accuracy and generalization ability of the IMPSO-BPNN model are the best overall. For example, the mean absolute percentage error (MAPE) of the optimal IMPSO-BPNN model (0.771%) is 5.85%, 12.62%, 17.96%, 7.57% and 7.88% less than that of the PSO-BPNN, GA-BPNN, BPNN, PSO-SVM and GA-SVM models (0.819%, 0.882%, 0.940%, 0.834% and 0.837%), respectively; and the correlation coefficient (R) is also higher than that of other models (0.99986). Secondly, the method shows visible superiority in the temporal dimension compared with the methods optimized by genetic algorithm and SVM method. For example, the CPU running time of the optimal IMPSO-BPNN model is reduced by 1773.32%, 149.74% and 1231.34% compared to that of the optimal GA-BPNN, PSO-SVM and GA-SVM models, respectively. Since the IMPSO-BPNN method is a flexible and general method, it is a new idea for the study of engine electronic control calibration tools.

Interaction force measurement of parallel robots based on structure-integrated force sensors using interpretable linear neural networks

The parallel robot in this research is mainly designed for the human-machine collaboration, which has two working modes: follow-up mode and pressing mode. In order to better realize these two functions, the measurement of the interaction force is mainly concerned. In the arrangement of the hardware, different from the six-dimensional force sensor method and current estimation method commonly used in industry, the structure-integrated force sensors (SIFS) measurement method is used. In the calculation method, a back propagation neural networks (BPNN) method is introduced to calculate the interaction force. Then, based on the prior knowledge of physics, a new interpretable linear neural networks (ILNN) method is proposed. The details of the ILNN algorithm and its differences from the classic neural network methods are given. Finally, the experimental results of the direct Newton–Euler method, the least square method, the BPNN method, and the ILNN method are compared. Experimental results show that the inference accuracy of the ILNN method is slightly better than the direct Newton–Euler method and the least square method, and its generalization ability is much better than the BPNN method. The use of the ILNN method based on the FIFS arrangement makes the measurement of the interaction force more accurate and reliable.

Radicalism Speech Detection in Indonesia on Twitter Using Backpropagation Neural Network Method

In this modern era, many people use social media easily and freely. One of the social media used is Twitter. The reason people use Twitter is that they can express their opinion freely. However, this freedom does not always have a positive impact on other Twitter users. One of the negative impacts for users is that they can spread radical content. Therefore, this research aims to detect whether a tweet contains radical elements or not using the backpropagation neural network method. The process is carried out by taking data on Twitter, after which the preprocessing process is carried out. Then the data is processed using imbalanced handling, where the data is divided into oversampling and undersampling data. After the data is divided, the next process is to do stopword and then look for accuracy by comparing different epoch values, namely 100, 150, 200, and 250. The best epoch value obtained is 200, with a final accuracy result of 86%.