public MalPositiveFeedbackRecResults GetRecommendations(MalPositiveFeedbackInput inputForUser, int numRecommendationsToTryToGet)
{
IUserInputClassifier <MalUserListEntries> classifier;
decimal targetScoreUsed;
if (inputForUser.TargetFraction != null)
{
classifier = new MalPercentageRatingClassifier(inputForUser.TargetFraction.Value, m_minEpisodesToClassifyIncomplete);
}
else
{
classifier = new MalMinimumScoreRatingClassifier(inputForUser.TargetScore.Value, m_minEpisodesToClassifyIncomplete);
}
IPositiveFeedbackForUser basicFeedback = inputForUser.AnimeList.AsPositiveFeedback(classifier,
additionalOkToRecommendPredicate: (animeId) => m_userCountByAnime.ContainsKey(animeId) && m_userCountByAnime[animeId] >= m_minUsersToCountAnime);
if (inputForUser.TargetFraction != null)
{
targetScoreUsed = basicFeedback.Items.Min(itemId => inputForUser.AnimeList.Entries[itemId].Rating ?? 10);
}
else
{
targetScoreUsed = inputForUser.TargetScore.Value;
}
IEnumerable <RatingPredictionRecommendation> recs = m_recommender.GetRecommendations(basicFeedback, numRecommendationsToTryToGet);
return(new MalPositiveFeedbackRecResults(recs, targetScoreUsed));
}
public IEnumerable <RatingPredictionRecommendation> GetRecommendations(IPositiveFeedbackForUser inputForUser, int numRecommendationsToTryToGet)
{
IList <int> userMediaLiteFeedback = new List <int>();
foreach (int realItemId in inputForUser.Items)
{
// Do not pass in items that MyMediaLite does not know about, it will crash
if (m_realItemIdToMediaLiteItemId.ContainsKey(realItemId))
{
int mediaLiteItemId = m_realItemIdToMediaLiteItemId[realItemId];
userMediaLiteFeedback.Add(mediaLiteItemId);
}
}
List <int> mediaLiteCandidateItemIds = m_realItemIdToMediaLiteItemId.Keys
.Where(realItemId => inputForUser.ItemIsOkToRecommend(realItemId))
.Select(realItemId => m_realItemIdToMediaLiteItemId[realItemId])
.ToList();
IList <Tuple <int, float> > mediaLiteScores = m_recommender.ScoreItems(userMediaLiteFeedback, mediaLiteCandidateItemIds);
List <RatingPredictionRecommendation> recs = new List <RatingPredictionRecommendation>();
foreach (Tuple <int, float> score in mediaLiteScores.OrderByDescending(p => p.Item2))
{
int mediaLiteItemId = score.Item1;
float predictedScore = score.Item2;
int realItemId = m_mediaLiteItemIdToRealItemId[mediaLiteItemId];
recs.Add(new RatingPredictionRecommendation(realItemId, predictedScore));
if (recs.Count >= numRecommendationsToTryToGet)
{
break;
}
}
return(recs);
}
public void Train(IBasicTrainingData <IPositiveFeedbackForUser> trainingData)
{
m_realUserIdToMediaLiteUserId = new Dictionary <int, int>();
m_mediaLiteUserIdToRealUserId = new Dictionary <int, int>();
m_nextMediaLiteUserId = 0;
m_realItemIdToMediaLiteItemId = new Dictionary <int, int>();
m_mediaLiteItemIdToRealItemId = new Dictionary <int, int>();
m_nextMediaLiteItemId = 0;
PosOnlyFeedback <SparseBooleanMatrix> mediaLiteFeedback = new PosOnlyFeedback <SparseBooleanMatrix>();
foreach (KeyValuePair <int, IPositiveFeedbackForUser> userFeedbackPair in trainingData.Users)
{
int userId = userFeedbackPair.Key;
IPositiveFeedbackForUser feedback = userFeedbackPair.Value;
m_realUserIdToMediaLiteUserId[userId] = m_nextMediaLiteUserId;
m_mediaLiteUserIdToRealUserId[m_nextMediaLiteUserId] = userId;
m_nextMediaLiteUserId++;
foreach (int itemId in feedback.Items)
{
if (!m_realItemIdToMediaLiteItemId.ContainsKey(itemId))
{
m_realItemIdToMediaLiteItemId[itemId] = m_nextMediaLiteItemId;
m_mediaLiteItemIdToRealItemId[m_nextMediaLiteItemId] = itemId;
m_nextMediaLiteItemId++;
}
mediaLiteFeedback.Add(m_realUserIdToMediaLiteUserId[userId], m_realItemIdToMediaLiteItemId[itemId]);
}
}
m_recommender.Feedback = mediaLiteFeedback;
m_recommender.Train();
}
