High Performance Computing Data: HPC
health prediction
The dataset contains ten variables, which are computer health readings for every 5-minute interval within 5
hours, including CPU load, fan speeds, memory usage, and power consumption. In total, we have 20 timesteps
of 467 nodes in the cluster. Our target variable for this data is the CPU temperature.
The dataset covers stock records for five weekdays each week,
in the period of 39 years,
from 1980 to 2019. Each
record contains the timestamp, stock price at “Open”,
“High”, “Low”, “Close”, and
“Volume” of the stock that
day. During the training and
testing process, we utilize
the attributes of the stock
price on Monday, Tuesday,
Wednesday, and Thursday to
predict Close price for Friday.
The dataset covers stock records for five weekdays each week,
in the period of 39 years,
from 1980 to 2019. Each
record contains the timestamp, stock price at “Open”,
“High”, “Low”, “Close”, and
“Volume” of the stock that
day. During the training and
testing process, we utilize
the attributes of the stock
price on Monday, Tuesday,
Wednesday, and Thursday to
predict Close price for Friday.