Thứ Hai, 21 tháng 12, 2015

Predicting how long to replace N servers node

My customer (a Data Center) give me a small dataset that contain time to replace N servers node.
I need to analysis and build predicting program that can predict how long to replace N servers node.

Input Training:

  •  Data of "p027.txt" show below:
Minutes Units
23    1 
29    2 
49    3 
64    4 
74    4 
87    5 
96    6 
97    6  
109   7  
119   8  
149   9  
145   9  
154   10  
166   10 

Example Input test:


2
40
8
10

Example Output:

Predicting minutes for units


35
622
128
159

Visualization about Model:  


### Code:
import graphlab
# if not csv format, need specify delimiter
sf = graphlab.SFrame.read_csv('http://www.ats.ucla.edu/stat/examples/chp/p027.txt',delimiter='\t')
graphlab.canvas.set_target('ipynb')
sf.show(view="Scatter Plot", x="Minutes", y="Units")

Finally, what i need:  SFrame - Python

Resources you will need:

  • Graphlab library
  • SFrame library

Example Code:


import graphlab
sf = graphlab.SFrame.read_csv('http://www.ats.ucla.edu/stat/examples/chp/p027.txt',delimiter='\t')
my_features1 = ["Units"]

my_features_model = graphlab.linear_regression.create(sf,target='Minutes',features=my_features1,validation_set=None)

repair_time = {
                           'Units':10
             }

print my_features_model.predict(repair_time)

Finish. Classify the problem to become linear regression single value. It's my simple way to solve.

Không có nhận xét nào:

Đăng nhận xét