Answer to Question #160716 in Algorithms for Habib

Question

1. There are huge secondary/ standard datasets published in the internet for such case studies:

·        Case 1: Text categorization: Spam Detection.

·        Case 2: Face detection.

·        Case 3: Signature recognition.

·        Case 4: Customer discover.

·        Case 5: Medicine: Classify if a patient has heart ischemia by a spectral analysis of his/her ECG.

 

Find one (1) of the secondary datasets from the internet. Name the secondary dataset that you have found.

 

1. Based on your answer in 1, perform the following computation using RapidMiner application software to compute accuracy performances (you have to screenshot the processes in detail):

 

1. Classification method: Artificial Neural Network (ANN) Algorithm.
2. Feature construction: Independent component analysis (ICA) method.
3. Feature selection: (Wrapper) method.
4. Results presentation in testing, training and validation: m-fold cross-validation (CV) method.

 

Answer 1

1(a) For spam detection we use SpamAssassin datasets. The text classification datasets are used to categorize the natural language text.

1(B) For the face detection, we use labelled faces in the wild(LFW) dataset.

1(c) For signature recognition, we use mainly three different kind of data set. which are CEDAR, UTSig, and ICDAR.

1(d) For customer dicover we use kaggale datasets.

1(e) ECG-Holter validated datasets is used for AI mechine learning algorithms validation.