COMPARATIVE PERFORMANCE FOR PREDICTIVE MODELLING IN MOTOR INSURANCE CLAIMS

2016 ◽  
Vol 78 (12-2) ◽  
Author(s):  
Zuriahati Mohd Yunos ◽  
Siti Mariyam Shamsuddin ◽  
Razana Alwee ◽  
Noriszura Ismail ◽  
Roselina Salleh@Sallehuddin
Keyword(s):  
Neural Network ◽  
Fuzzy Inference ◽  
Back Propagation ◽  
Predictive Modelling ◽  
Back Propagation Neural Network ◽  
Insurance Claims ◽  
Inference System ◽  
Neuro Fuzzy ◽  
Motor Insurance ◽  
Claim Frequency

The expected claim frequency and the expected claim severity are used in predictive modelling motor insurance claims. There are two categories of claims were considered, namely, third party property damage and own damage. Datasets from the year 2001 to 2003 are used to develop the predictive model. This paper proposes three different methods, namely, regression analysis, back propagation neural network and adaptive neuro fuzzy inference system to model claim frequency and claim severity as the two important elements in modelling the motor insurance claims. The experimental results showed that the back propagation neural network model produces more accurate as compared to regression analysis and adaptive neuro fuzzy inference system in predicting the claim frequency and claim severity. For both OD and TPPD claim, the results have shown the lowest MAPE with 0.2191 and 0.6515, and 0.2169 and 0.326, respectively.

scholarly journals ANALISA DAN PERBANDINGAN AKURASI MODEL PREDIKSI RENTET WAKTU ARUS LALU LINTAS JANGKA PENDEK

2015 ◽  
Vol 6 (3) ◽  
pp. 148
Author(s):  
Bambang Lareno
Keyword(s):  
Neural Network ◽  
Fuzzy Inference ◽  
Back Propagation ◽  
Back Propagation Neural Network ◽  
Percentage Error ◽  
Mean Absolute Deviation ◽  
Mean Square ◽  
Absolute Deviation ◽  
Inference System ◽  
Neuro Fuzzy

<p>Abstrak <br />Terdapat banyak algoritma yang dapat dipakai untuk memprediksi arus lalu lintas, namun belum diketahui algoritma manakah yang memiliki kinerja lebih akurat untuk lalu lintas di Indonesia. Algoritma-algoritma tersebut perlu diuji untuk mengetahui algoritma manakah yang memiliki kinerja lebih akurat. Metode yang diusulkan adalah metode perbandingan tingkat akurasi dari algoritma berbasis neural network yang bisa digunakan untuk prediksi data rentet waktu. Algoritma yang akan diuji adalah back Propagation Neural Network (BP-NN), Adaptive Neuro Fuzzy Inference System (ANFIS), Wavelet Neural Network (WNN), dan Evolving Neural Network (ENN), yang digunakan untuk memprediksi arus lalulintas. Masing-masing algoritma akan implementasikan dengan menggunakan MatLab 2009b. Pengukuran kinerja dilakukan dengan menghitung rata-rata error yang terjadi melalui besaran Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) dan Mean Absolute Deviation (MAD). Semakin kecil nilai dari masing-masing parameter kinerja ini menyatakan semakin dekat nilai prediksi dengan nilai sebenarnya. Dalam penelitian ini diketahui bahwa Algoritma ENN memprediksi arus lalu lintas dengan lebih akurat.</p>

scholarly journals ANFIS AND NEURAL NETWORK BASED FACIAL EXPRESSION RECOGNITION USING CURVELET FEATURES

2014 ◽  
pp. 255-261
Author(s):  
S.P. Khandait ◽  
R.C. Thool ◽  
P.D Khandait
Keyword(s):  
Neural Network ◽  
Facial Expression ◽  
Facial Expression Recognition ◽  
Fuzzy Inference ◽  
Back Propagation ◽  
Curvelet Transform ◽  
Back Propagation Neural Network ◽  
Expression Recognition ◽  
Inference System ◽  
Promising Tool

Curvelet transform is a promising tool for multi-resolution analysis on images. This paper explains a new approach for facial expression recognition based on curvelet features extracted using curvelet transform. Curvelet transform is applied on the database images and curvelet coefficients are obtained for selected scale for image analysis. Facial curvelet features are compressed using singular value decomposition (SVD) approach. Back propagation neural network (BPNN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) are used as classifiers for classifying expressions into one of the seven categories like angry, disgust, fear, happy, neutral, sad and surprise. Experimentation is carried out on JAFFE database. The experimental results show that the novel approach is a better option for extracting feature values and classifying facial expressions.

An ANFIS Based Model for Predicting Frost Heaving in Seasonal Frozen Regions

2014 ◽  
Vol 501-504 ◽  
pp. 391-394
Author(s):  
Yi Ming Xiang ◽  
Xue Yan Liu ◽  
Gui Xiang Ling ◽  
Bin Du
Keyword(s):  
Clustering Algorithm ◽  
Fuzzy Inference ◽  
Learning Algorithm ◽  
Back Propagation ◽  
Back Propagation Neural Network ◽  
Subtractive Clustering ◽  
Forecast Errors ◽  
Frost Heaving ◽  
Inference System ◽  
Neuro Fuzzy

An adaptive neuro-fuzzy inference system (ANFIS) model has been developed to predict frost heaving in seasonal frozen regions. The structure of ANFIS is initialized by the subtractive clustering algorithm. The hybrid learning algorithm consisting of back-propagation and least-squares estimation is used to adjust parameters of ANFIS and automatically produce fuzzy rules. The data of frost heaving test obtained from a literature are used to train and check the system. The predicted results show that the proposed model outperforms the back propagation neural network (BPNN) in terms of computational speed, forecast errors, and efficiency. The ANFIS based model proves to be an effective approach to achieve both high accuracy and less computational complexity for predicting frost heaving.

Optimized neural network and adaptive neuro-fuzzy controlled dynamic voltage restorer for power quality performance

2021 ◽  
Vol 0 (0) ◽  
Author(s):  
Prashant Kumar ◽  
Sabha Raj Arya ◽  
Khyati D. Mistry
Keyword(s):  
Neural Network ◽  
Power Distribution ◽  
Fuzzy Inference ◽  
Dynamic Performance ◽  
Hybrid Approach ◽  
Voltage Regulation ◽  
Dynamic Voltage Restorer ◽  
Inference System ◽  
Intelligence System ◽  
Neuro Fuzzy

Abstract In this article, a hybrid approach is implemented namely, neural network training (NNT) based machine learning (ML) estimator inspired by artificial neural network (ANN) and self-adaptive neuro-fuzzy inference system (ANFIS) to tackle the voltage aggravations in the power distribution network (DN). In this work, potential of swarm intelligence technique namely particle swam optimization (PSO) is analysed to obtain an optimum prediction model with certain modifications in training algorithm parameters. In practice, when the systems are continuously subjected to parametric changes or external disturbances, then ample time is dedicated to tune the system to regain its stable performance. To improve the dynamic performance of the system intelligence-based techniques are proposed to overcome the shortcomings of conventional controllers. So, gain tuning process based on the intelligence system is a desirable choice. The statistical tools are used to proclaim the effectiveness of the controllers. The obtained MSE, RMSE, ME, SD and R were evaluated as 0.0015959, 0.039949, −0.00089838, 0.039941 and 1 in the training phase and 0.0015372, 0.039207, −0.0005657, 0.039203 and 1 in the testing phase, respectively. The results revealed that the ANFIS-PSO network model could accomplish a better DC voltage regulation performance when it is compared to the conventional PI. The proposed intelligence strategies confirm that the predicted DVR model based on NNT-ML and ANFIS has faster convergence speed and reliable prediction rate. Moreover, the simulation results show that the dynamic response is improved with proposed PSO based NNT based ML and ANFIS (Takagi-Sugeno) that significantly compensates the voltage based PQ issues. The proposed DVR is actualized in MATLAB/SIMULINK platform.

Model Identification of 3R Palnar Robot using Neural Network and Adaptive Neuro-Fuzzy Inference System

2021 ◽  
Author(s):  
R. Subasri ◽  
R. Meenakumari ◽  
R. Velnath ◽  
Srinivethaa Pongiannan ◽  
M. Sri Sai Mani Rohit Kumar
Keyword(s):  
Neural Network ◽  
Fuzzy Inference System ◽  
Fuzzy Inference ◽  
Model Identification ◽  
Inference System ◽  
Neuro Fuzzy

Comparison of different modeling methods toward predictive capability evaluation of the bonding strength of wood laminated products

2021 ◽  
pp. 095440892110530
Author(s):  
Morteza Nazerian ◽  
Seyed Ali Razavi ◽  
Ali Partovinia ◽  
Elham Vatankhah ◽  
Zahra Razmpour
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Response Surface Methodology ◽  
Response Surface ◽  
Bonding Strength ◽  
Fuzzy Inference ◽  
Molar Ratio ◽  
Inference System ◽  
Neuro Fuzzy ◽  
Artificial Neural

The main aim of this study is usability evaluation of different approaches, including response surface methodoloy, adaptive neuro-fuzzy inference system, and artificial neural network models to predict and evaluate the bonding strength of glued laminated timber (glulam) manufactured using walnut wood layers and a natural adhesive (oxidized starch adhesive), with respect to this fact that using the modified starch can decrease the formaldehyde emission. In this survey, four variables taken as the input data include the molar ratio of formaldehyde to urea (1.12–1.52), nanocellulose content (0%–4%, based on the dry weight of the adhesive), the mass ratio of the oxidized starch adhesive to the urea formaldehyde resin (30:70–70:30), and the press time (4–8 min). In order to find the best predictive performance of each selected evaluation approach, different membership functions were used. The optimal results were obtained when the molar ratio, nanocellulose content, mass ratio of the oxidised starch, and press time were set at 1.22, 3%, 70:30, and 7 min, respectively. Based on the performance criteria including root mean square error (RMSE) and mean absolute percentage error (MAPE) obtained from the modeling of response surface methodology, adaptive neuro-fuzzy inference network, and artificial neural network, it became evident that response surface methodology could offer a better prediction of the response with the lowest level of errors. Beside, artificial neural network and adaptive neuro-fuzzy inference system support the response surface methodology approach to evaluate bonding strength response with high precision as well as to determine the optimal point in fabrication of laminated products.

Sign in / Sign up

Export Citation Format

Share Document