d04904cdd1df7afba9859868f4e487d3a2332070,demo_semiotic.py,,,#,6

Before Change

newsgroups_train = fetch_20newsgroups(subset="train", remove=("headers", "footers", "quotes"))

vectorizer = TfidfVectorizer()
tfidf_X = vectorizer.fit_transform(newsgroups_train.data)

corpus = st.CorpusFromScikit(
	X=CountVectorizer(vocabulary=vectorizer.vocabulary_).fit_transform(newsgroups_train.data),
	y=newsgroups_train.target,
	feature_vocabulary=vectorizer.vocabulary_,
	category_names=newsgroups_train.target_names,
	raw_texts=newsgroups_train.data
).build()

semiotic_square = st.SemioticSquare(
	corpus,
	category_a="alt.atheism",
	category_b="soc.religion.christian",
	neutral_categories=["talk.religion.misc"]
)

html = st.SemioticSquareViz(semiotic_square).get_html(num_terms=5)

fn = "demo_semiotic_square_atheism_christianity.html"
open(fn, "wb").write(html.encode("utf-8"))
print("Open " + fn + " in Chrome or Firefox.")

After Change

	scorer=st.LogOddsRatioUninformativeDirichletPrior(alpha_w=0.001)
)

html = st.produce_semiotic_square_explorer(semiotic_square,
                                           jitter=0.01,
                                           category_name="Fresh",
                                           not_category_name="Rotten",
                                           x_label="Rotten-Fresh",
                                           y_label="Plot-Review",
                                           neutral_category_name="Plot Description",
                                           metadata=movie_df["movie_name"],
                                           x_axis_values=[-2.58, -1.96, 0, 1.96, 2.58],
                                           y_axis_values=[-2.58, -1.96, 0, 1.96, 2.58])

fn = "demo_semiotic.html"
open(fn, "wb").write(html.encode("utf-8"))
print("Open " + fn + " in Chrome or Firefox.")

Instances