internal DataModelSimilaritiesIterator(UserSimilarity otherSimilarity, long[] itemIDs)
{
this.otherSimilarity = otherSimilarity;
this.itemIDs = itemIDs;
this.i = 0;
this.itemID1 = itemIDs[0];
this.j = 1;
}
public GenericUserSimilarity(UserSimilarity otherSimilarity, DataModel dataModel, int maxToKeep)
{
this.similarityMaps = new FastByIDMap <FastByIDMap <double> >();
long[] itemIDs = longIteratorToList(dataModel.getUserIDs());
IEnumerator <UserUserSimilarity> allSimilarities = new DataModelSimilaritiesIterator(otherSimilarity, itemIDs);
this.initSimilarityMaps(TopItems.getTopUserUserSimilarities(maxToKeep, allSimilarities).GetEnumerator());
}
public NearestNUserNeighborhood(int n, double minSimilarity, UserSimilarity userSimilarity, DataModel dataModel, double samplingRate)
: base(userSimilarity, dataModel, samplingRate)
{
int num = dataModel.getNumUsers();
this.n = (n > num) ? num : n;
this.minSimilarity = minSimilarity;
}
public AbstractUserNeighborhood(UserSimilarity userSimilarity, DataModel dataModel, double samplingRate)
{
this.userSimilarity = userSimilarity;
this.dataModel = dataModel;
this.samplingRate = samplingRate;
this.refreshHelper = new RefreshHelper(null);
this.refreshHelper.addDependency(this.dataModel);
this.refreshHelper.addDependency(this.userSimilarity);
}
public CachingUserSimilarity(UserSimilarity similarity, int maxCacheSize)
{
Action refreshRunnable = null;
this.similarity = similarity;
this.similarityCache = new Cache <Tuple <long, long>, double>(new SimilarityRetriever(similarity), maxCacheSize);
if (refreshRunnable == null)
{
refreshRunnable = () => this.similarityCache.clear();
}
this.refreshHelper = new RefreshHelper(refreshRunnable);
this.refreshHelper.addDependency(similarity);
}
/**
* <p>
* Constructs a based on the given {@link UserSimilarity}. By
* setting {@code samplingRate} to a value less than 1.0, this implementation will only examine that
* fraction of all user-user similarities between two clusters, increasing performance at the expense of
* accuracy.
* </p>
*/
public FarthestNeighborClusterSimilarity(UserSimilarity similarity, double samplingRate)
{
if (similarity == null)
{
throw new Exception("similarity is null");
}
//Preconditions.checkArgument(similarity != null, "similarity is null");
//Preconditions.checkArgument(!Double.IsNaN(samplingRate) && samplingRate > 0.0 && samplingRate <= 1.0,
// "samplingRate is invalid: %.4f", samplingRate);
if (!(!Double.IsNaN(samplingRate) && samplingRate > 0.0 && samplingRate <= 1.0))
{
throw new Exception("samplingRate is invalid");
}
this.similarity = similarity;
this.samplingRate = samplingRate;
}
public GenericUserBasedRecommender(DataModel dataModel, UserNeighborhood neighborhood, UserSimilarity similarity)
: base(dataModel)
{
Action refreshRunnable = null;
this.neighborhood = neighborhood;
this.similarity = similarity;
if (refreshRunnable == null)
{
refreshRunnable = () => this.capper = this.buildCapper();
}
this.refreshHelper = new RefreshHelper(refreshRunnable);
this.refreshHelper.addDependency(dataModel);
this.refreshHelper.addDependency(similarity);
this.refreshHelper.addDependency(neighborhood);
this.capper = this.buildCapper();
}
protected override float doEstimatePreference(long theUserID, long[] theNeighborhood, long itemID)
{
if (theNeighborhood.Length == 0)
{
return(float.NaN);
}
DataModel model = this.getDataModel();
UserSimilarity similarity = base.getSimilarity();
float num = 0f;
bool flag = false;
foreach (long num2 in theNeighborhood)
{
if ((num2 != theUserID) && model.getPreferenceValue(num2, itemID).HasValue)
{
flag = true;
num += (float)similarity.userSimilarity(theUserID, num2);
}
}
return(flag ? num : float.NaN);
}
public override long[] getUserNeighborhood(long userID)
{
DataModel model = this.getDataModel();
FastIDSet set = new FastIDSet();
IEnumerator <long> enumerator = SamplingLongPrimitiveIterator.maybeWrapIterator(model.getUserIDs(), this.getSamplingRate());
UserSimilarity similarity = this.getUserSimilarity();
while (enumerator.MoveNext())
{
long current = enumerator.Current;
if (userID != current)
{
double d = similarity.userSimilarity(userID, current);
if (!(double.IsNaN(d) || (d < this.threshold)))
{
set.add(current);
}
}
}
return(set.toArray());
}
//零数据情况下建立用户相似度数据集
public void CountSimilarity()
{
IEnumerable <User> users = db.Users;
foreach (var user in users)
{
IList <SimilarityMap> smilarityMap = new List <SimilarityMap>();
UserExtensionInfo currentuserExtensionInfo = db.UserExtensionInfos.Single(x => x.UserId == user.Id);
IEnumerable <int> otherUsersId = SelectOtherUsers(user);
foreach (var otherUserid in otherUsersId) //其他用户
{
UserExtensionInfo otherExtensionInfo = db.UserExtensionInfos.Single(x => x.UserId == otherUserid);
int similarityDegree = AcquireSimilarityBetween(currentuserExtensionInfo, otherExtensionInfo);
SimilarityMap newsimilarityMap = new SimilarityMap
{
UserId = otherUserid,
Similarity = similarityDegree
};
smilarityMap.Add(newsimilarityMap);
}
IEnumerable <SimilarityMap> similarity = smilarityMap.OrderByDescending(x => x.Similarity).Take(SimilarityUserSettingCount);
int priority = 1;
foreach (var similarityMap in similarity)
{
UserSimilarity userSimilarity = new UserSimilarity
{
UserId = user.Id,
SimilarUserId = similarityMap.UserId,
Priority = priority
};
db.UserSimilarities.Add(userSimilarity);
priority++;
}
}
db.SaveChanges();
}
public ThresholdUserNeighborhood(double threshold, UserSimilarity userSimilarity, DataModel dataModel, double samplingRate)
: base(userSimilarity, dataModel, samplingRate)
{
this.threshold = threshold;
}
internal SimilarityRetriever(UserSimilarity similarity)
{
this.similarity = similarity;
}
internal Estimator(UserSimilarity userSimilarityImpl, long theUserID, double minSim)
{
this.userSimilarityImpl = userSimilarityImpl;
this.theUserID = theUserID;
this.minSim = minSim;
}
public CachingUserSimilarity(UserSimilarity similarity, DataModel dataModel)
: this(similarity, dataModel.getNumUsers())
{
}
internal MostSimilarEstimator(long toUserID, UserSimilarity similarity, Rescorer <Tuple <long, long> > rescorer)
{
this.toUserID = toUserID;
this.similarity = similarity;
this.rescorer = rescorer;
}
public NearestNUserNeighborhood(int n, double minSimilarity, UserSimilarity userSimilarity, DataModel dataModel)
: this(n, minSimilarity, userSimilarity, dataModel, 1.0)
{
}
public NearestNUserNeighborhood(int n, UserSimilarity userSimilarity, DataModel dataModel)
: this(n, double.NegativeInfinity, userSimilarity, dataModel, 1.0)
{
}
/**
* <p>
* Constructs a based on the given {@link UserSimilarity}. All
* user-user similarities are examined.
* </p>
*/
public FarthestNeighborClusterSimilarity(UserSimilarity similarity)
: this(similarity, 1.0)
{
}
public GenericUserSimilarity(UserSimilarity otherSimilarity, DataModel dataModel)
{
this.similarityMaps = new FastByIDMap <FastByIDMap <double> >();
long[] itemIDs = longIteratorToList(dataModel.getUserIDs());
this.initSimilarityMaps(new DataModelSimilaritiesIterator(otherSimilarity, itemIDs));
}
public GenericBooleanPrefUserBasedRecommender(DataModel dataModel, UserNeighborhood neighborhood, UserSimilarity similarity)
: base(dataModel, neighborhood, similarity)
{
}
public ThresholdUserNeighborhood(double threshold, UserSimilarity userSimilarity, DataModel dataModel)
: this(threshold, userSimilarity, dataModel, 1.0)
{
}
