The hybrid systems and technologies have always attracted scientist and researchers to pursue their inventions and research. More than 10000 research papers on hybrid technologies are available in search results in digital search libraries. In this paper, we are focusing on the hybrid system where one of the components is ANN. The paper has two phases. In the first phased we are presenting a literature review of hybrid systems based on ANN. In the second phase, we are discussing the concept of hybrid neural expert system. At the end, we are exploring the possibilities for development of hybrid neural expert system for evaluating motivational strategies on human resources.

1) Expert system for feature for speaker identification using lip feature with PCA (mehta A, 2010)

The above paper presents comparative analysis for speaker identification by using lip features, Principal component analysis(PCA) and neural network classifiers. The input are the height of the outer corners of the mouth, width of the outer corners of mouth, height of the inner corners of the mouth, width of the inner corners of the mouth, height of the upper lip and height of the lower lip. Back propagation network is used for learning. Maximum of 91.07% accuracy has been obtained. Financial Analysis

2) Case base reasoning with feature weights derived by BP Network (Yan Peng, 2007)

Case based reasoning is the method used for decision making in the continuous changing environment. The system is hybrid combination of case base reasoning along with back propagation algorithm. This system is used for fault detection and diagnosis. Higher performance is obtained with hybrid system compared to normal convention case base reasoning methods.

3) Intelligent system for diagnosis of Epilepsy (Shukla A, 2009)

Epilepsy is a chronic neurological disorder. The paper presents intelligent diagnosis system for Epilepsy. The system is developed using ANN and Neuro Fuzzy technique. The system use feed forward neural network and trained using back propagation algorithm. The system also have adaptive inference engine. The system normalized the value of every attribute of input lies between zero and one. Total 265 instances have been collected. Out of which 200 were used for training the system and 65 have been used for testing the system. The MATLAB has been used for developing simulation. The results have been compared for accuracy, training time, number of neurons and number of epochs. The system has shown improved performance over the other system.

4) An Intelligent system for false alarm reduction in infrared forest fire detection (Begona C. Arrue, 2000)

The system combines neural network and expert fuzzy rules to detect forest fires in open areas. The back propagation algorithm is used for training the system. Fuzzy expert rule base is used to express knowledge in rules. The study presented indicates the dramatic decrease in the false alarm rate maintaining the detection capabilities.

5) Application of Artificial Neural Network in Sales forecasting (W.W.H. Yip, 1997)

The system forecast the sales volume for the future. The choice of network topologies were Silva’s adaptive back propagation algorithm and the network architecture were selected by genetics algorithm. The network can forecast in advance up to six months of sales. The test result of this system is compared with time series method of forecasting. The result shows that ANN system is better.

6) Rule generation from neural network for student assessment (McAlister M.J., 1999)

The above system has been implemented with the use of back propagation algorithm. The system is considered as hybrid system, because it extracts rules from the weight stored in neurons. After extracting the rules the same system can be developed as a rule base expert system.

7) Fault diagnosis in power system-substation level-through hybrid artificial neural network and expert system (El-Fergany A.A, 2011)

This system is developed using combination of ANN and expert system for offline fault diagnosis in power system. The system use information of the operated relays and tripped circuit breakers. With the hybrid system, multiple faults are also identified. The system also identify the most probable faulted section based on performance indices. The proposed ANN identify the faulted section even if the fault is in the inured zone. The communication problems in the relay signals either missing or noisy can also be detected. Expert system module used the rule base technique. The comparison between ANN, ES and the hybrid intelligent system for fault diagnosis is presented as well.

The neural network is tested using Stuttgart Neural Network simulator on Linux platform and the code generated in C/C++ can is portable and can be complied on any windows based system. Expert system has been developed using prolog.

8) Design of an expert system to estimate cost in an automated job shop manufacturing system (Fazlollahtabar H, 2010)

The proposed cost estimation model is based on fuzzy rule back propagation network. The main objective is to estimate the cost under uncertainty. A multiple linear regression analysis is applied to identify rules. Numeric example showed the effectiveness of proposed model for cost estimation under uncertainty.

9) Comprehensive evaluation of logistics systems efficiency based on the integrated DEA/BPN model (Weihual Xia, 2009)

The proposed system combines management science with AI. The system evaluates the operational efficiency of its every subsystem for logistics system management. The result of the system shows that the average testing error of the integrated DEA/BPN Model is reduced to 89% in comparison with that of traditional BPN model.

Analysis of Literature Review

The literature review presented above is summarised in the following table for the analysis purpose. The analysis has been done based on the following factors. Domain, Objective of the system, hybrid techniques, Comparison criteria, and performance.

Table 1: Analysis of Hybrid Systems Based on Parameters
Table1Analysis of Hybrid System with-1


1. The above analysis reveals that hybrid systems with ANN has one of its components are developed in different domains like social science, mechanical, HR, Sales, medical and manufacturing.

2. When ANN is combined with the other technique, the system has shown the better performance.

3. Another factor which can be noticed is that MATLAB is also very preferred tool to develop hybrid system based on ANN.

The use of MATLAB is also justified in the following research paper.

Research on IDSS by MATLAB integration: Theory and method (Gao, 2011)

The paper presents the integration MATLAB and Intelligent DSS. The paper here concludes that MATLAB will improve efficiency and provide practical way for developing ANN Based Solution. MATLAB will also reduce the number of lines of code.