Question
Answer the following questions:1 A data mining routine has been applied to a transaction dataset and has classified 88 records as fraudulent (30 correctly so) and 952 as non- fraudulent (920 correctly so). Construct the classification matrix and calculate the error rate, sensitivity, and specificity: 2. Suppose that this routine has an adjustable cutoff (threshold) mechanism by which you can alter the proportion of records classified as fraudulent: Describe how moving the cutoff up or down woul
Answer the following questions: 1 A data mining routine has been applied to a transaction dataset and has classified 88 records as fraudulent (30 correctly so) and 952 as non- fraudulent (920 correctly so). Construct the classification matrix and calculate the error rate, sensitivity, and specificity: 2. Suppose that this routine has an adjustable cutoff (threshold) mechanism by which you can alter the proportion of records classified as fraudulent: Describe how moving the cutoff up or down would affect Sensitivity, specificity, true negative, and true positive:
Answers
Answer the following questions: 1
A data mining routine has been applied to a transaction dataset and has classified 88 records as fraudulent (30 correctly so) and 952 as non- fraudulent (920 correctly so). Construct the classification matrix and calculate the error rate, sensitivity, and specificity: 2. Suppose that this routine has an adjustable cutoff (threshold) mechanism by which you can alter the proportion of records classified as fraudulent: Describe how moving the cutoff up or down would affect Sensitivity, specificity, true negative, and true positive:
Okay with this problem. Um They're basically asking us to use a set of data to compare standard deviation and I. Q. R. When we change a data value and they go back and use the data from question number 25. Um where we were using cell phone usage and fraud. And so um the first thing they ask us to do is they ask us to use that data and go back and calculate standard deviation and like you are for the original set of data and I used my T I. 84 to do that. So the standard deviation um was just about 58. And then the I. Q. R. For this original set of data was 4 89.5. That's Q three minus Q. One which is 4 33. So that gives us an I. Q. R. Of 56.5. And that was minutes we were seeing how many minutes they spoke a month. Okay. And then on the second part um we were asked again to calculate standard deviation in I. Q. R. But it asked us to take the month that they spoke 346 minutes and change it to zero and think about how it's gonna affect standard deviation in I. Q. R. Now, think about what we're doing, we're taking 300 almost 50 minutes and we're turning it into zero. And so when we recalculate our stats, look at what happens. Standard deviation when that number become zero is right about 115 like it doubles compared to the 58. And then the I. Q. R. With our new data is 4 89.5 minus 4. 33 were using Q. Three and Q one to calculate I Q. R. 56.5. So our I. Q. R. Stays exactly the same. And what they're really wanting you to think about is we've talked about the term resistant Yeah. Whether a measure is resistant to extreme values or not, and what you want to realize is standard deviation is not resistant because that change had a great impact on the standard deviation. But like you are is resistant. So even though I changed that 3 52 0, the I Q R. Didn't change at all. Yeah. Yeah.
Hi there. In this question, we are asked to calculate the percent error for each of the three trials for students. See from the table. All right, so we need the equation for percent. Air percent error is calculated by taking the error. So that was obtained. By taking the value that was obtained and subtracting, he accepted value and notice that has absolute value signs around it. So that will always be positive. And then we divide that by the accepted value what it should have been and then to change that decimal to a percent, we multiplied by 100. Right? So this equation we're going to follow. Let's go ahead and compute trial. What percent air for trial? One? Yeah, is equal to 0.11 grams per centimeter cubed. And the accepted value for this density is 1.59 grams per centimeter. Cubed. We need multiple that by 100. So we get a percent error of 6.9%. He had a rounded that to two significant figures since 20.11 on Lee has two significant figures. Okay, let's move on to trial. Too percent error. The error this time that were given in the table is 0.10 grams per centimeter cubed, divided by the accepted value of 1.59 g per centimeter. Cute multiplied by 100. I lost 20 here, and that gives us 6.3%. All right, And finally trial three. Same drill here using the same equation. Except now in trial three, the error was calculated to be 0.12 grams per centimeter cubed, dividing that by 1.59 times. I'm sorry. 1.59 g per centimeter. Cute. Okay. Times 100 gives us a percent of 7.5%. All right, so there we go. We have calculated the percent error for each of the three trials. Trial one was 6.9% error. Try trial to with 6.3% error. And trial three was 7.5% hair. Thanks so much for watching
Came today and I'll be gone over questioning their 32 trans Lou talking person here. First it's stringers leading saw. The heirs are for trial. One trying to try all three. You're the air wells. Make it in way 19. Okay? Because the furniture you like tea this year here to was less than only three waas my point for So in order to calculate the percent here you do the absolute one undermined by your no, by 100 you. So it doesn't point person here. Trial one Carol choose by 0.6 Trump three is a equipment.
We look at the data set, there seems to be ah, variety of different pikes, and because of that, the heights appear to be actually measured because if it was reported, there would be repeated heights.