Geography Reference
In-Depth Information
tile size can be a very valuable input of the neural network to correctly classify the cachability
of requests.
During the training process, a training record corresponding to the request of a particular tile
is associated with a boolean target (0 or 1) which indicates whether the same tile is requested
again or not in window, as shown in Equation 5.
1
if the tile is requested again in window
0
otherwise
target
=
(5)
Once trained, the neural network output will be a real value in the range [0,1] that must be
interpreted as the probability of receiving a successive request of the same tile within the time
window. A request is classified as
cacheable
if the output of the neural network is above 0.5.
Otherwise, it is classified as
non cacheable
.
The neural network is trained through supervised learning using the data sets from the
extracted trace files. The trace data is subdivided into training, validation, and test sets, with
the 70%, 15% and 15% of the total requests, respectivelly. The first one is used for training the
neural network. The second one is used to validate that the network is generalizing correctly
and to identify overfitting. The final one is used as a completely independent test of network
generalization.
Each training record consists of an input vector of recency, frequency and size values, and the
known target. The weights are adjusted using the backpropagation algorithm, which employs
the gradient descent to attempt to minimize the squared error between the network output
values and the target values for these outputs [36]. The network is trained in batch mode, in
which weights and biases are only updated after all the inputs and targets are presented. The
pocket algorithm, which saves the best weights found in the validation set, is used.
Neural network performance is measured by the correct classification ratio (CCR), which
computes the percentage of correctly classified requests versus the total number of processed
requests.
CartoCiudad
IDEE-Base
training
76.5952
75.6529
validation
70.2000
77.5333
test
72.7422
82.7867
Table 5.
Correct classification ratios (%) during training, validation and testing for Cartociudad and
IDEE-Base.
Figure 16 shows the CCRs obtained during training, validation and test phases for
Cartociudad and IDEE-Base services. As can be seen, the neural network is able to correctly
classify the cachability of requests, with CCR values over the testing data set ranging between
72% and 97%, as shown in Table 5. The network is stabilized to an acceptable CCR within 100
to 500 epochs.
