The code requires about 1.9GB to run through. Please ensure satisfy this memory requirement. And due to the implement of parallel computing,
this code perfers to be run on a multicores machine which all cpus are free.
Please use:
cd src
javac -cp . PearsonsCorrelation.java
java PearsonsCorrelation -trainingFile <trainingFile> -outputFile <outputFile>
to get matrix file. (The commind is the same as provied initially).Please fill in the trainingFile and outputFile
Or you can use the makefile just by changing traininFile and outputFile.
The parameters has already been set in the code.
1)The number of threads are 4.
2)The minimal number of commonly rated movies----MinCom is 15
If number of commonly rated movies is less than MinCom, the Persons correlation will be Set to NaN.
Please use:
cd src
javac -cp . MovieRunner.java
java MovieRunner -trainingFile <trainingFile> -testFile <testFile> -matrixFile <matrixFile>
to get evaluation result. (The commind is the same as provied initially).Please fill in the trainingFile , testFile and matrixfile
Or you can use the makefile just by changing traininFile, testFile and matrixFile
The parameters has already been set in the code.
1)The maximum number of neighbors to predict ratings----MaxNei is infinite
If the number of neighbors can be used to predict ratings, we will only choose top MaxNei neighbors to make prediction.