Reference no: EM132514918
DATA MINING PROJECT
Select the leaf data set for our data mining project
Business and data understanding
Leaf data set is multivariate data. It includes 40 different plant spices. It contains 340 instances and 16 attribute. Name of included attributes are given below
• Class(Species)
• Specimen Number
• Eccentricity
• Aspect Ratio
• Elongation
• Solidity
• Stochastic Convexity
• Isoperimetric Factor
• Maximal Indentation Depth
• Lobedness
• Average Intensity
• Average Contrast
• Smoothness
• Third moment
• Uniformity
• Entropy
In these attributes different plants are classified into 36 classes and specimen numbers represents that how many spacemen produced by the plant according to the class whereas rest of attributes explains the characteristics of plant. The data type of all attributes is real. All values of attributes are available in this data set. Associated task of given data set is classification. We classify the class of different plants according to the characteristics of leaf.
Modelling
According to the CRIS-DM model next phase is modelling. In this phase we select the operator that will be useful to do the classification of class attribute. We make an effort to apply the four different operators such as decision tree, Random Forest Tree, Deep Learning, and Naive Bayes. Then, Naive Bayes are unsuccessful for leaf data set because the data is in numerical form and Naive Bayes cannot handle the numerical data. As a result, we select decision tree, random forest tree, and deep learning operator to use in the process.
Next, we select the root mean squared error criteria to analyse the result. In addition to this, we select other useful operators to complete the process such as select attribute, set role, split data, apply model, and performance.
Select attribute operator: we used this operator to select the attributes for our process. This operator allows selecting which attributes should be a part of resulting.
Set role operator: This operator to set the role of attribute and target the role as label.
Split data: This operator split up the data into two or more examples sets or separates the data.
Apply model: This model is trained by another operator to the example set. The goal of this model is to predict or classify invisible data and transform the data by applying a pre-processing model.
Performance: This operator is used for statistical performance evaluation of classification tasks. This operator provides a list of performance criteria for the classification task.
Evaluation
This is the evaluation phase. In this phase, we use the operators in the Rapid miner software for the objective.
1. Decision tree: A decision tree is a set of trees, similar to set of nodes, designed to create decisions about classifying values into classes or estimating numerical target values. Each node represents a split rule for a particular attribute. This operator can handle an example set containing nominal and numeric properties. In this process, firstly import the data.
Attachment:- Data mining project.rar