Closed gqqnbig closed 4 years ago
电影年份是连续的,可以用GaussianNB,但流派是离散的,应该用BernoulliNB。按https://stackoverflow.com/questions/14254203/mixing-categorial-and-continuous-data-in-naive-bayes-classifier-using-scikit-lea 说的方法试试看。
准确率:0.54
我还是修改了一下naive_bays.py。
fit的时候删除'userId', 'movieId',准确率提升到0.5486859960857063
对每个用户创建model,准确率提升到0.557498282776833
你对naive_bays_combine.py作同样修改试试?
电影年份是连续的,可以用GaussianNB,但流派是离散的,应该用BernoulliNB。按https://stackoverflow.com/questions/14254203/mixing-categorial-and-continuous-data-in-naive-bayes-classifier-using-scikit-lea 说的方法试试看。