Exam 14: Data Mining

The primary advantage of neural networks is that they provide more accurate predictions,especially when the relationships are linear.

The K-Means clustering algorithm :

Today's organizations rely on their quantitative experts,who have access to large amounts of data,to make sense of it in a timely manner.

The last step of cluster analysis is to understand the shared characteristics of the observations in each cluster.​

The function INDEX(A1:E5,4,3)would return the value in cell C4.​

Exhibit 14.1 The age, weight, gender, and the number of children were tracked for 50 people who either have or have not been diagnosed with a certain illness.  The data are provided in the table below. 31 43 177 4 32 42 178 1 33 28 183 1 34 60 177 1 35 60 176 5 36 62 199 0 37 57 177 0 38 31 152 3 39 68 185 2 40 26 134 2 41 38 142 2 42 51 199 2 43 62 133 4 44 38 131 4 45 73 196 1 46 48 120 2 47 54 197 1 48 38 136 0 49 54 183 0 50 52 199 0 -See Exhibit 14-1 Use neural nets via Palisade's NeuralTools to classify the people as being diagnosed as having an illness or not. (Reserve 20% of the observations for testing.) How well does this method correctly classify people

Which of the following is false regarding neural networks ​

Exhibit 14.1 The age, weight, gender, and the number of children were tracked for 50 people who either have or have not been diagnosed with a certain illness.  The data are provided in the table below. 31 43 177 4 32 42 178 1 33 28 183 1 34 60 177 1 35 60 176 5 36 62 199 0 37 57 177 0 38 31 152 3 39 68 185 2 40 26 134 2 41 38 142 2 42 51 199 2 43 62 133 4 44 38 131 4 45 73 196 1 46 48 120 2 47 54 197 1 48 38 136 0 49 54 183 0 50 52 199 0 -See Exhibit 14.1 - What kind of person is more likely to be diagnosed with the illness

The logistic function 1/(1+e ^-x )

Exhibit 14-3 Information for 26 colleges and universities in the state of Indiana is provided in the table below.  The following questions contain a sequence of steps for executing the K-Means Clustering Method. ​  -Refer to Exhibit 14-3 Choose any three college indices from 1 to 26 as trial values for the cluster centers. Complete the logicto return the name of the college or university associated with the index as well as the standardized numeric measures.​

Which of the following is false concerning a data warehouse

Exhibit 14.1 The age, weight, gender, and the number of children were tracked for 50 people who either have or have not been diagnosed with a certain illness.  The data are provided in the table below. 31 43 177 4 32 42 178 1 33 28 183 1 34 60 177 1 35 60 176 5 36 62 199 0 37 57 177 0 38 31 152 3 39 68 185 2 40 26 134 2 41 38 142 2 42 51 199 2 43 62 133 4 44 38 131 4 45 73 196 1 46 48 120 2 47 54 197 1 48 38 136 0 49 54 183 0 50 52 199 0 -See Exhibit 14.1 - How likely would a 40 year old,200 pound father of 2 children be to have the illness

Exhibit 14.2 The age, weight, gender, and the number of children were tracked for 25 people.  It is not known if these people have or have not been diagnosed with a certain illness.  The data are provided in the table below. person age weight gender children illness 51 52 86 m 2 52 72 145 f 2 53 33 113 f 2 54 60 167 m 0 55 65 134 f 1 56 71 198 m 2 57 26 120 f 1 58 63 165 m 2 59 23 155 f 1 60 52 152 f 1 61 28 183 m 3 62 59 177 m 1 63 48 184 m 2 64 51 201 m 0 65 71 181 m 2 66 26 179 m 0 67 34 115 f 1 68 56 176 m 1 69 65 126 m 2 70 62 181 f 1 71 49 157 m 1 72 71 165 f 0 73 21 138 f 2 74 51 187 m 0 75 21 119 f 1 -See Exhibit 14-2 Using Palisade's NeuralTools and the results from the prior problem,use the prediction data in this table to classify these 25 people as having been diagnosed with an illness or not.

Exhibit 14.1 The age, weight, gender, and the number of children were tracked for 50 people who either have or have not been diagnosed with a certain illness.  The data are provided in the table below. 31 43 177 4 32 42 178 1 33 28 183 1 34 60 177 1 35 60 176 5 36 62 199 0 37 57 177 0 38 31 152 3 39 68 185 2 40 26 134 2 41 38 142 2 42 51 199 2 43 62 133 4 44 38 131 4 45 73 196 1 46 48 120 2 47 54 197 1 48 38 136 0 49 54 183 0 50 52 199 0 -See Exhibit 14.1 - How well can logistic regression classify the people as having the illness or not

Better classification methods should improve lift,which is the increase in purchases gained through marketing to people with the highest probability of purchasing.

For logistic regression:

In data partitioning,data can be divided into training,testing,and prediction subsets.

Exhibit 14-3 Information for 26 colleges and universities in the state of Indiana is provided in the table below.  The following questions contain a sequence of steps for executing the K-Means Clustering Method. ​  -Refer to exhibit 14-3 As a first step in grouping the Indiana colleges and universities into clusters,standardize the numeric measures.

Which of the following is true of clustering methods ​

In the classification method,when data are partitioned into the training and testing subsets: