def __init__(self):
self.persistModels = {
self.KNN_MODEL_KEY: KNNmodel(),
self.MP_MODEL_KEY: MostPopularModel(),
self.CL_MODEL_KEY: ClusteringModel(),
self.CF_MODEL_KEY: CFmodel()
}
self.transientModels = {self.SI_MODEL_KEY: {}}
def __init__(self):
# offline models saved in a dictionary, key: static variable defined above, value: instance of class as attribute
self.myModels = {
self.MP_MODEL_KEY: MostPopularModel(),
self.RP_MODEL_KEY: RecentPopularModel(),
self.KNN_MODEL_KEY: KNNmodel(),
self.CL_MODEL_KEY: ClusteringModel(),
self.CF_MODEL_KEY: CFmodel()
}
def __init__(self):
self.persistModels = {
self.KNN_MODEL_KEY: KNNmodel(),
self.MP_MODEL_KEY: MostPopularModel(),
self.CL_MODEL_KEY: ClusteringModel(),
self.CF_MODEL_KEY: CFmodel()
}
# similarity model is used for each user
self.transientModels = {self.SI_MODEL_KEY: {}}
def __init__(self):
self.persistModels = {
self.KNN_MODEL_KEY: KNNmodel(),
self.MP_MODEL_KEY: MostPopularModel(),
self.CL_MODEL_KEY: ClusteringModel(),
self.CF_MODEL_KEY: CFmodel()
}
# similarity model is used for each user
# online recommendation, trained by online learner
self.transientModels = {self.SI_MODEL_KEY: {}}
else:
self.recs = []
def predict(self, itemFeature):
# X should be item's category feature, only single record
# return the similar items
itemFeature = itemFeature.values.reshape(1, -1)
center, indices = self.clusteringModel.predict(itemFeature)
return indices[0]
def provideRec(self):
return self.recs
if __name__ == "__main__":
from DatabaseInterface import DatabaseInterface
from Models.ClusteringModel import ClusteringModel
db = DatabaseInterface("../DATA")
db.startEngine()
itemFeatureTable = db.extract(
DatabaseInterface.ITEM_FEATURE_KEY).loc[:, "unknown":]
model = ClusteringModel()
model.train(itemFeatureTable)
modelSI = SimilarItemModel(model)
modelSI.train(itemFeatureTable.loc[1], 4)
print(modelSI.provideRec())
modelSI.train(itemFeatureTable.loc[1], 2)
print(modelSI.provideRec())
else:
self.recs = []
def predict(self, itemFeature):
# X should be item's category feature, only single record
# return the similar items
itemFeature = itemFeature.values.reshape(1, -1)
center, indices = self.clusteringModel.predict(itemFeature)
return indices[0]
def provideRec(self):
return self.recs
if __name__ == "__main__":
from DatabaseInterface import DatabaseInterface
from Models.ClusteringModel import ClusteringModel
db = DatabaseInterface("../DATA")
db.startEngine()
itemFeatureTable = db.extract(DatabaseInterface.ITEM_FEATURE_KEY).loc[:, "unknown":]
model = ClusteringModel()
model.train(itemFeatureTable)
modelSI = SimilarItemModel(model)
modelSI.train(itemFeatureTable.loc[1], 4)
print modelSI.provideRec()
modelSI.train(itemFeatureTable.loc[1], 2)
print modelSI.provideRec()