*: The combination of probabilistic reasoning, fuzzy logic, neural network and evolutionary computation 

*: Hard computing relies on predefined instructions like a numerical analysis and brisk software and uses two-valued logic. Soft computing is based on the model of the human mind where it has probabilistic reasoning, fuzzy logic, and uses multivalued logic

Q2
*:The combination of rule extraction component along with neural network enables neural expert system to justify and provide explanation facility for its conclusion. Neural network also allows dealing with noisy and incomplete data because of its capability of generalisation.

*:

QUESTION 1          
- A hybrid intelligent system is one that combines at least two intelligent technologies.

- The combination of probabilistic reasoning, fuzzy  logic, neural networks, and evolutionary  computation forms the core of soft computing,

- Soft computing is based on the model of the human mind where it has probabilistic reasoning and fuzzy logic.
- Hard computing relies on binary logic and predefined instructions like numerical analysis and software.

QUESTION 2
- The use of a rule extraction component in combination with a neural network allows the neural expert system to justify and explain its conclusions. Because of their ability to generalise, neural networks can also deal with noisy and incomplete input. As a result, approximation reasoning is allowed.

QUESTION 3
- When dealing with raw data, neural networks are low-level computing structures that function well. Fuzzy logic is involved with higher-level reasoning based on linguistic data gathered from domain expertise.

Summary of ANFIS

ANFIS uses a hybrid learning algorithm that combines the least-square estimator and the gradient descent method. Neuron output is calculated on a layer-by-layer basis, and the rule consequent parameters are identified. Each epoch is composed from a forward pass and a backward pass.Researchers have discovered that the ANFIS modelling technique can be utilised as a feasible alternative to conducting experiments while doing analysis.

- Hybrid intelligence system is combination of at least 2 intelligent technology. For example combination of neural network and fuzzy system.

- Soft computing exploits the tolerance for uncertainly and imprecision to achieve greater tractability and robustness 

- Hard computing fails to produce any solution while soft computing still capable of finding good solution

Question 2

Neural expert system capable of approximate reasoning because it is inference engine that controls the data flow in the system and initiates inference over knowledge network. The input data does not have to precisely match the data that was used.

 

Question 3

Fuzzy system and neural network considered to be natural complementary because neural network can perform well when dealing with the raw data because it is low level computation structures and fuzzy logic deal with reasoning on higher level using linguistic information.

ANFIS
An ANFIS apply hybrid learning algorithm that combine the least-squares estimator and the gradient descent method. In training algorithm each epoch composed form forward pass and backward pass.  In forward pass, training set of input pattern is presented to ANFIS neuron outputs are calculated on the layer by layer and rule consequent parameter are identified.  In backwards pass the back propagation algorithm is applied the error signals are propagated back and antecedent parameters are updated according to chain rule. 

Hard Computing is that the ancient approach employed in computing that desires Associate in Nursing accurately declared analytical model.

Question 2
The combination of the rule extraction component and the neural network enables the neural expert system to justify and explain its conclusion. Because of its ability to generalise, neural networks can also deal with noisy and incomplete data. As a result, it allows for approximate reasoning.

Question 3
Fuzzy logic and neural networks are natural complements in the development of intelligent systems. While neural networks are low-level computational structures that perform well with raw data, fuzzy logic deals with higher-level reasoning by utilising linguistic information obtained from domain experts.

An adaptive neuro-fuzzy inference system, also known as an adaptive network-based fuzzy inference system (ANFIS), is a type of artificial neural network based on the Takagi–Sugeno fuzzy inference system. It has the potential to capture the benefits of both neural networks and fuzzy logic principles because it integrates both in a single framework. Its inference system is represented by a set of fuzzy IF–THEN rules with the ability to learn and approximate nonlinear functions. As a result, ANFIS is regarded as a universal estimator. The best parameters obtained by genetic algorithm can be used to use the ANFIS in a more efficient and optimal manner.

A hybrid intelligent system is one that combines at least two intelligent technologies. For example, combining a neural network with a fuzzy system results in a hybrid neuro-fuzzy system. The constitutes of the core of soft computing is the combination of probabilistic reasoning, fuzzy logic, neural networks and evolutionary. The differences between har and soft computing is when hard computing fails to produce any solution, soft computing is still capable of finding good solutions.

QUESTION 2: Why is a neural expert system capable of approximate reasoning? Draw a neural knowledge base for a three-class classification problem. Suppose that an object to be classified is either an apple, an orange or a lemon.

Because it has neural inference engine where it controls the information flow in the system and initiates inference over the neural knowledge base.

QUESTION 3: Why are fuzzy systems and neural networks considered to be natural complementary tools for building intelligent systems? Draw a neuro-fuzzy system corresponding to the Sugeno fuzzy inference model for the implementation of the AND operation. Assume that the system has two inputs and one output, and each of them is represented by two linguistic values: small and large.

Because while neural networks are low-level computational structures that perform well when dealing with raw data, fuzzy logic deals with reasoning on a higher level, using linguistic information acquired from domain experts.

 

Core of soft computing is the combination of probabilistic reasoning, fuzzy logic, neural networks and evolutionary computation. 

 

Differences between hard and soft computing is that hard computing is not able to produce solutions whereas soft computing are able to find good solutions very well. Other than that, hard computing is by binary logic while soft computing is based of a model of human mind. 

 

Q2- Neural expert capable of approximate reasoning because the information flow in the system is controlled and the initiates inference over the neural knowledge base. 

 

Q3- Fuzzy systems and neural networks considered to be natural complementary tools for building intelligent systems because fuzzy logic is high level reasoning and using linguistic information acquired from domain experts. 

- Neural expert system is capable of approximate reasoning because it allows dealing with noisy and incomplete data because of its capability of generalization.

Question 3

- Fuzzy systems and neural networks are considered to be natural complementary tools for building intelligent systems because they are natural complementary tools in building intelligent systems. Neural networks are low-level computational structures that perform well when dealing with raw data. Meanwhile fuzzy logic deals with reasoning on a higher level, using linguistic information acquired from domain experts.

- The neuro-fuzzy system corresponds to a fuzzy model of Takagi–Sugeno, wherein the weights of the ANN model are similar to the parameters of the fuzzy system. This structure is called ANFIS. It is a hybrid model composed of a fuzzy and artificial neural network. The ANFIS structure consists of five layers, namely, fuzzy layer, product layer, normalized layer, de-fuzzy layer, and total output layer.