wk-7

Running Head: DATA MINING 1

DATA MINING 2

Title

Student’s name

Instructor

Course

Date

DATA MINING

List and briefly describe the nine-step process in conducting a neural network project.

A neural network is an artificial neural network that is composed of artificial nodes. In a nutshell, we can say that this is machine learning capable of modeling itself, creating an artificial neural network (Sharda, Delen, & Turban, 2020). The development process involves nine steps as listed below:

Step 1; Data collection

This stage involves the collection of data that will e applied during the existing and training. This stage is very critical, and it is important to ensure that the issue in question has adequate data.

Step 2: Training and test data separation.

This step involves the identification o the data that had been collected in the earlier step. The data is also separated, and there is a validation of the networks that will be used. Using numerous data increases the processing time to improve data accuracy.

Step3; Decision making in network architecture

In this step, one chooses the best network architecture o choice. The learning method selection is performed as well as the network architecture that will be used. The choice of the applied neural network is dependent n the capacity of the staff available as well as the available development tools

Step4; selecting a learning algorithm

It is important to consider the success story of the algorithm. This includes the feedback multilayer neural networks efficacy in prediction. The other aspect of considering is the layers and the number of neurons where, in some cases, the network design is performed through the use of genetic algorithms.

Step5; Setting parameters and initializing the weight

For the learning performance that is desired, some settings are made in adjusting the network level. This step also includes determining the duration o the training; the initial values are used as they are transformed into feedback on performance training while setting weights and network settings.

Step 6: Data processing

For accuracy of data results, representation and order of application influence the data efficacy.

Step 7and 8. Training and testing

The input data and the desired output are represented in the ANN when learning is carried out interactively. The calculated output is supposed to be within the acceptable tolerance when the ANN calculates output and adjusts the weights

Step 9: implementation

This phase involves obtaining a stable set of weight. For the training set, the network is supposed to reproduce the desired output according to the input.

What is the main difference between classification and clustering? Explain using concrete examples. Add a few Examples.

Classification and clustering are two types of learning methods used in data mining for data analysis and dividing data based on particular classification rules of association between the objects. Classification and clustering may appear similar, but they are two different processes based on their meanings. Clustering helps group physical objects into classes of similar objects based on the principle of intraclass similarity and decreasing interclass similarity (Sharda, Delen, & Turban, 2020). Clustering in the search engine helps display the results containing the key specified term in the search box. For instance, when a word is inserted in the engine box, it should produce that keyword results.

On the other hand, classification is the process of finding a set of functions describing and distinguishing data classes through using the function in predicting the class of the object whose class is already known. Classification analyzes class-labeled data objects. For example, by using a decision tree, t is possible to classify algorithms on a data set whose lust of research is known.

The other major difference between clustering and classification is that the former is an unsupervised learning technique, which groups similar instances based on features. This means that there is only one set of input data for this type of algorithm that is required to obtain information without knowing the output. The latter is a supervised learning technique. This means that we know the input, and we know the possible output o the algorithm. 

References

Sharda, R., Delen, D., Turban, E. (2020). Analytics, data science, & artificial intelligence: systems for decision support, global edition. pearson education limited. Retrieved from: https://www.pearson.com/us/higher-education/program/Sharda-Analytics-Data-Science-Artificial-Intelligence-Systems-for-Decision-Support-11th-Edition/PGM2067063.html