def kdata():
print("\nK-Dataset")
w2v = dict(size=512, negative=100, window=3)
run_context_graph(KDATA,
ismat=True,
thresh=1,
wl=10,
nw=200,
w2v_params=w2v)
def reuters():
print('\nReuters')
w2v = dict(size=256, negative=100, window=3)
run_context_graph(REUTERS,
thresh=5,
wl=1,
nw=800,
top_n=10,
w2v_params=w2v)
def webkb4():
print('\nWebKB-4')
w2v = dict(size=512, negative=50, window=3)
# Best thresh 5 so far
run_context_graph(
WEBKB,
thresh=5,
wl=1,
nw=800,
top_n=4,
w2v_params=w2v,
ignore_list=[
'"other"'
] # Only counts "entity representing" categories in benchmark
)
def webkb():
print('\nWebKB')
w2v = dict(size=512, negative=50, window=3)
# Best thresh 10 so far
run_context_graph(WEBKB, thresh=10, wl=1, nw=800, w2v_params=w2v)
def cstr():
print('\nCSTR')
w2v = dict(size=128, negative=5, window=3)
run_context_graph(CSTR, thresh=1, nw=200, wl=10, w2v_params=w2v)