There are four nodes in the SAS Enterprise Miner
time-series data mining work process:
File Import – The File Import node allows user to upload
and convert external flat files, spreadsheets, and
database table into format that SAS Enterprise Miner
can recognize as a data source and use it in the
subsequent data mining processes.
TS Data Preparation – The TS Data Preparation (TSDP)
node converts the input data into time-series data for
analysis. A few settings were set for this research
analysis. Firstly, the timeseries column, which contains
the time interval in the input OD Time Interval format,
was set to the role of Time ID in this analysis. The Time
ID would form up the x-axis in the generated time-series
data plots. The passenger column, which contains the
frequency of number of passengers, was set to the role of
Target. The Target would form up the y-axis in the
generated time-series data plots. As we are interested to
examine the passenger volume of each MRT train
station, we will set the Origin column, which contain the
origin train station ID, as the cross-sectional variable,
Cross ID.
Metadata – The Metadata node allow users to modify
certain data attributes so that the data is suitably
formatted for the next process node.
TS Similarity – The TS Similarity (TSS) node performs
the clustering and similarity analysis by comparing the
time-series and group time-series that exhibit similar
characteristics over time. As the time series might have
different lengths, DTW technique will be applied to
compare two time-series; the input and target sequences.
The TSS node also calculates the similarity measures
between the compared input and target sequences. The
Result function of the TSS node visualizes the results of
the similarity and clustering analysis.
There are four nodes in the SAS Enterprise Miner
time-series data mining work process:
File Import – The File Import node allows user to upload
and convert external flat files, spreadsheets, and
database table into format that SAS Enterprise Miner
can recognize as a data source and use it in the
subsequent data mining processes.
TS Data Preparation – The TS Data Preparation (TSDP)
node converts the input data into time-series data for
analysis. A few settings were set for this research
analysis. Firstly, the timeseries column, which contains
the time interval in the input OD Time Interval format,
was set to the role of Time ID in this analysis. The Time
ID would form up the x-axis in the generated time-series
data plots. The passenger column, which contains the
frequency of number of passengers, was set to the role of
Target. The Target would form up the y-axis in the
generated time-series data plots. As we are interested to
examine the passenger volume of each MRT train
station, we will set the Origin column, which contain the
origin train station ID, as the cross-sectional variable,
Cross ID.
Metadata – The Metadata node allow users to modify
certain data attributes so that the data is suitably
formatted for the next process node.
TS Similarity – The TS Similarity (TSS) node performs
the clustering and similarity analysis by comparing the
time-series and group time-series that exhibit similar
characteristics over time. As the time series might have
different lengths, DTW technique will be applied to
compare two time-series; the input and target sequences.
The TSS node also calculates the similarity measures
between the compared input and target sequences. The
Result function of the TSS node visualizes the results of
the similarity and clustering analysis.
การแปล กรุณารอสักครู่..
