Using XLMINER’s neural network routine to fit a model using XLMINER default valu
ID: 665658 • Letter: U
Question
Using XLMINER’s neural network routine to fit a model using XLMINER default values for neural network parameters by using the predictors such as CRIM, ZN, INDUS, CHAS, NOX, RM, AGE, DIS, RAD to classify the value of CAT.MEDV.
i. Record the RMS errors for the training data and the validation data, and observe the lift charts for repeating the process, changing the number of epochs to 300, 3000, 10,000, 20,000.
ii. What happens to RMS error for the training data set as the number of epochs increases?
iii. What happens to RMS error for the validation data set as the number of epochs increases?
iv. Comments on the appropriate number of epochs for the model.
Explanation / Answer
RMS error:
The RMS error is also known as root-mean-squared error which calculates the square root of the average squared error. It determines the typical error for original data which comprises of same scale of the original data.
(i)
The prediction data set consists of the variables CRIM, ZN, INDUS, CHAS, NOX, RM, AGE, DIS and RAD. An additional variable CAT.MEDV is a categorical variable formed by created by dividing the median MEDV values into high and low values.
If number of epochs are increased by 300, the RMS error for CRIM= 13.
If number of epochs are increased by 3000, the RMS error for CRIM= 130.
If number of epochs are increased by 10000, the RMS error for CRIM= 417.
If number of epochs are increased by 20000, the RMS error for CRIM= 834.
The RMS error for the remaining data variables are calculated in the same way.
Thus, by observing the values, a conclusion can be made that the RMS errors increases as number of epochs are increased.
(ii)
When the number epochs increase over the training data set it is very difficult to fit the error in the training data. Because, the RMS error values depends on different training data set values.
(iii)
In general the validation data is used to overcome the overfitting of the data. When the number of epochs increases for the validation data set, then the RMS error or testing error decreases.
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.