def window_diversity(window, n_topics, iters, thresh):
artist_count = window.groupby(["artist"])["user"].count().reset_index()
window = window[window["artist"].\
isin(artist_count[artist_count["user"]>=thresh]["artist"])]
ui_matrix = window[["user", "artist", "sum"]]\
.pivot(index = "user", columns = "artist", values = "sum")
users = ui_matrix.index
ui_matrix = ui_matrix.fillna(0)
ui_matrix_tfidf = np.multiply(tf, idf)
ui_lda = LatentDirichletAllocation(n_topics=n_topics,
max_iter=iters,
learning_method='online',
learning_offset=50.,
random_state=0).fit(ui_matrix_tfidf)
ti_matrix = ui_lda.components_
ut_matrix = ui_lda.fit_transform(ui_matrix_tfidf)
diversity = hf.LDADiversity(ut_matrix)
user_diversity = pd.DataFrame({"diversity": diversity}, index=users)
return (user_diversity)
def get_recommendations(window_data, n_users=None, weights=None):
entropy = hf.LDADiversity(window_data["it"])
popularity = (window_data["ui"] > 0).sum(axis=0)
users = window_data["users"]
if weights == None:
weights = np.array([10, 1, 1])
if n_users == None:
n_users = len(users)
r_weighted_all = {}
for i in range(n_users):
user = users[i]
current_artists_idx = np.where(window_data["ui"].loc[user] > 0)[0]
sim_rec = top_sim(window_data["artists"], current_artists_idx,
window_data["artist_sim"])
div_rec = top_div(window_data["artists"], current_artists_idx, entropy,
popularity.values)
clust_rec = top_clustering(window_data["artists"], current_artists_idx,
window_data["artist_sim"])
rec_sum = pd.merge(sim_rec, div_rec, on="artist")
rec_sum = pd.merge(rec_sum, clust_rec, on="artist")
r_weighted = weighted_recommendations(rec_sum, weights)
r_weighted_all[user] = r_weighted
return (r_weighted_all)
iu_matrix_tfidf = np.multiply(tf, idf)
#transform item_user matrix to item_topic and topic_user matrix
iu_lda = LatentDirichletAllocation(n_topics=n_topics,
max_iter=iters,
learning_method='online',
learning_offset=50.,
random_state=0).fit(iu_matrix_tfidf)
tu_matrix = iu_lda.components_
it_matrix = iu_lda.fit_transform(iu_matrix_tfidf)
user_sim = hf.LDASim(ut_matrix)
item_sim = hf.LDASim(it_matrix)
user_div = hf.LDADiversity(ut_matrix)
hf.LDAClustering(item_sim, ui_matrix)
hf.draw_graph(graph)
#Clustering bipartite graph
#forming bipartite graph W
m, n = ui_matrix.shape
axis_label = np.concatenate((ui_matrix.index, ui_matrix.columns))
W = np.block([[np.zeros((m, m)), ui_matrix], [ui_matrix.T, np.zeros((n, n))]])
D = np.diag(W.sum(axis=0))
W_scaled = np.linalg.inv(D)**(1 / 2) @ W @ np.linalg.inv(D)**(1 / 2)
u, s, v = np.linalg.svd(W_scaled)
x = np.linalg.inv(D)**(1 / 2) @ u[:, 1]
def track_diversity(window_data):
entropy = hf.LDADiversity(window_data["ut"])
entropy_labeled = dict(zip(window_data["users"], entropy))
return (entropy_labeled)
