gsm1011 / fall-2010 Goto Github PK

It is my fault, but remember every time you write a program, do a systematic 
test and produce a report to make sure that every point is correctly 
implemented and tested. 

Data mining projects are data and computing intensive ones, so it is important 
to keep to as efficient as you can.

In project one, my program is a little slower than other people's program, I 
think the main reason lies in that I failed to sorted each gene data during the 
entropy calculation process, which will try all possible splits and find the 
one with the highest information gain. So it is advisable that gene data be 
sorted before being processed. 

For project two, efficiency problem will be more critical, because if not 
correctly processed, it will take longer time to process. So, be careful with 
the algorithm implementations. 

When creating a new object by using the new operator, always try to test if the 
memory allocation is successful or not. Two methods can be used for this 
purposes. 
  # use the std::nothrow to depress throwing exceptions. 
  # use the try {} catch(...) {} programming block so that the thrown exception objects can be caught. 

Annotations have been added to the related file. 

Some itemset with lower than minimum support are put into the output.