PDF Publication Title:
Text from PDF Page: 177
Table 3. Contingency table for predictions when employing different training and test data sets. Finally, to prove that synthetic and original data are similar, but not identical, we subtract (element-wise) the original and synthetic datasets, calling the resulting dataset as difference- dataset. If the similarity between the datasets is true, the values of the difference-dataset should have zero mean (no reproduction bias) and a relatively small standard deviation (not too small to make both datasets indistinguishable or too large to make them completely unrelated). Table 4 shows the mean and standard deviation for 20 continuous and discrete features of the above- mentioned difference-dataset. We can observe, in both cases, that the behavior is the expected one. The synthetic features exhibit variability (no exact copy) and are centered on expected values (no bias). Similarly, Figure 8 gives the distribution of values for several continuous and discrete features (upper and lower diagrams, respectively) of the difference-dataset. The Y-axis of the histograms corresponds to the number of repetitions of a given value, and the X-axis represents the values of the difference-dataset. These values are, in all cases, in the interval [- 1,1], due to the [0,1] scaling and one-hot encoding of the continuous and discrete features, respectively (Section 3.1). Table 4. Mean and standard deviation of difference between values (original vs. synthetic) for several features. Doctoral Thesis: Novel applications of Machine Learning to NTAP - 175PDF Image | Novel applications of Machine Learning to Network Traffic Analysis
PDF Search Title:
Novel applications of Machine Learning to Network Traffic AnalysisOriginal File Name Searched:
456453_1175348.pdfDIY PDF Search: Google It | Yahoo | Bing
Cruise Ship Reviews | Luxury Resort | Jet | Yacht | and Travel Tech More Info
Cruising Review Topics and Articles More Info
Software based on Filemaker for the travel industry More Info
The Burgenstock Resort: Reviews on CruisingReview website... More Info
Resort Reviews: World Class resorts... More Info
The Riffelalp Resort: Reviews on CruisingReview website... More Info
CONTACT TEL: 608-238-6001 Email: greg@cruisingreview.com (Standard Web Page)