System for Prediction of Non Stationary Time Series based on the Wavelet Radial Bases Function Neural Network Model

Kusdarwati, Heni; Handoyo, Samingun

doi:10.11591/ijece.v8i4.pp2327-2337

Cited by 9 publications

(8 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…e output gate determines the output content of LSTM, and the calculation is shown in equations ( 5) and (6). First, the sigmoid function is used to determine the cell content to be output, and then the tanh function is used to convert the cell state value between −1 and 1 to obtain the final output:…”

Section: Deep Learning Models and Prediction Algorithmsmentioning

confidence: 99%

“…Reference [ 5 ] used multiple financial indicators as eigenvalues of modeling data for the four types of audit opinions issued by companies and established a multicategory audit opinion prediction model based on error correction output coding and support vector machine (SVM). Reference [ 6 ] established a two-category prediction model of bank loan risk level classification authenticity audit based on SVM. However, most models for forecasting audit opinions have only been developed in the context of individual financial statements, and none of the models dealt with consolidated financial statements.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Intelligent Financial Auditing Model Based on Deep Learning

Xiao-feng

Zhu

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

The entire auditing process is complicated and tedious and requires a lot of human resources. Therefore, the intelligent development of auditing is the general trend. In order to improve the audit quality, this paper establishes an intelligent financial audit model that can predict the audit opinion of the consolidated financial statements. This paper proposes an audit opinion prediction model based on the fusion of deep belief neural network (DBN) and long-short term memory (LSTM). First, an indicator system is established for audit opinions, and multiple financial parameters are used to describe possible audit opinions. On this basis, a DBN network is designed to complete deep feature extraction and used for LSTM training. According to the prediction model obtained by training, the subsequent audit opinion can be scientifically predicted. In the experiment, the method in this paper is tested based on financial audit related data sets and compared with the prediction results of traditional multilayer perceptron (MLP), convolutional neural network (CNN), and LSTM models. The results verify the validity and reliability of the model in this paper.

show abstract

Section: Deep Learning Models and Prediction Algorithmsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Intelligent Financial Auditing Model Based on Deep Learning

Xiao-feng

Zhu

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…is type of method models tra c ow through nite parameters and does not depend on the size of the dataset. Reference [8] analyzed the network tra c of many cellular base stations and distinguished the tra c into two parts, i.e., predictable and unpredictable ones, which proves that the predictable tra c has autocorrelation. Reference [9] proposed a seasonal SARIMA model, which accurately captured the seasonal characteristics of network tra c by analyzing the autocorrelation of time series, and then obtained longterm tra c forecast results.…”

Section: Introductionmentioning

confidence: 99%

5G Traffic Prediction Based on Deep Learning

Gao

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

The demand of wireless access users is increasing explosively. The 5G network traffic is increasing exponentially and showing a trend of diversity and heterogeneity, which makes network traffic forecasting face many challenges. By studying the actual performance of the 5G network, this paper makes an accurate prediction of the 5G network and builds a smoothed long short-term memory (SLSTM) traffic prediction model. The model updates the number of layers and hidden units according to the prediction accuracy adaptive mechanism. At the same time, in order to reduce the randomness of the 5G traffic sequence, the output feature sequence of the original time series is stabilized by the seasonal time difference method. In the experiments, the prediction results of the proposed algorithm are compared with those of the traditional algorithms. The results show that the SLSTM algorithm can effectively improve the accuracy of 5G traffic prediction. The model can be used for 5G traffic prediction for decision-making.

show abstract

“…The prediction of a numerical value in machine learning is conducted by regression models. Some deployments of regression models for forecasting time series data are carried out by Kusdarwati and Handoyo [9] predicting nonstationary time series using a hybrid model between wavelet and neural networks, Handoyo and Marji [10] predicting the EURO to IDR exchange rate using the fuzzy least square, and Handoyo et al [11] predicting multiple time series of regional wages using a hybrid model between fuzzy system and fuzzy c-mean.…”

Section: Introductionmentioning

confidence: 99%

Finding the Best Performance of Bayesian and Naïve Bayes Models in Fraudulent Firms Classification through Varying Threshold

Widodo¹,

Handoyo²,

Irianto³

et al. 2022

IJETAE

View full text Add to dashboard Cite

—Fraud detection is the first step to preventing fraud committed by both individuals and organizations. The development of a high-performance classification model to detect fraud is an interesting topic in machine learning modeling. A finding of the best Bayesian and Naive Bayes classification models is a crucial issue because both models are simple and easily applied models in the fields of life and social sciences. This study aims to obtain the best performance of classification models developed based on probability concepts, namely Bayesian and Naive Bayes models. Adding a threshold value to the decision-making criteria of the two models is an effort expected to can find models that perform superiorly. Data on the auditing of fraudulent firms containing of 775 firms from various business sectors in Australia is used as a case study. The testing data consisting of 100 instances were taken by cluster random sampling with a proportion of 61 non-fraudulent and 39 fraudulent firms and the remaining instances as the training data. The best Bayesian model has an average accuracy of 84% obtained at a threshold value of 0.22. While the best Naive Bayes model has an average accuracy of 94% which is obtained in the 15 threshold values. Adding the threshold value has a significant impact on the performance of the Bayesian model, which can increase the average accuracy from 36% to 84%. On the other hand, the average accuracy of the Naive Bayes model only increased by 1%, from 93% to 94%. Performance measures Sensitivity, Specificity, F1 score, and ROC curve of the Naive Bayes model are also superior to the Bayesian model.

show abstract

System for Prediction of Non Stationary Time Series based on the Wavelet Radial Bases Function Neural Network Model

Cited by 9 publications

References 17 publications

Intelligent Financial Auditing Model Based on Deep Learning

Intelligent Financial Auditing Model Based on Deep Learning

5G Traffic Prediction Based on Deep Learning

Finding the Best Performance of Bayesian and Naïve Bayes Models in Fraudulent Firms Classification through Varying Threshold

Contact Info

Product

Resources

About