我正在创建用于股票市场分析的情绪分析代码。这是代码的核心:
import flair
flair_sentiment = flair.models.TextClassifier.load('en-sentiment')
columns = ['ticker', 'date', 'time', 'headline']
parsed_and_scored_news = pd.DataFrame(parsed_news, columns=columns)
sentiment = []
for head in parsed_and_scored_news['headline']:
s = flair.data.Sentence(head)
flair_sentiment.predict(s)
total_sentiment = s.labels
sentiment.append(total_sentiment)
scores_df = pd.DataFrame(sentiment)
parsed_and_scored_news = parsed_and_scored_news.join(scores_df, rsuffix='_right')
# Convert the date column from string to datetime
parsed_and_scored_news['date'] = pd.to_datetime(parsed_and_scored_news.date).dt.dateparsed_and_scored_news.head()
产生以下输出:
ticker date time headline 0
0 AMZN 2021-03-26 02:37PM Tech stocks are going to do vey well going for... POSITIVE (0.9986)
1 AMZN 2021-03-26 01:17PM Amazon mocked idea its drivers urinated in bot... NEGATIVE (0.9855)
2 AMZN 2021-03-26 01:11PM ThredUp CEO on IPO day: Dont tax resale and Am... NEGATIVE (0.6743)
3 AMZN 2021-03-26 12:54PM Why this retailer is seeing a triple-digit sal... POSITIVE (0.9597)
4 AMZN 2021-03-26 12:07PM How to secure your smart home camera POSITIVE (0.9981)
因为我想将数据输入到机器学习模型中,所以我需要分数是数字。我知道使用 probability = sentence.labels[0].score
只给我们分数,但这意味着没有办法将一个陈述是肯定的还是否定的进行分类。有没有办法在归类为负面的分数后面添加一个“-”(否定)符号。例如 - NEGATIVE (0.9855) = -9855
。这将确保信息既是数字又是有用的。
答案 0 :(得分:0)
这段代码对我有用:
sentiment = []
sentiment_score =[]
for head in parsed_and_scored_news['headline']:
s = flair.data.Sentence(head)
flair_sentiment.predict(s)
total_sentiment = s.labels[0].value
total_sentiment_score = s.labels[0].score
sentiment.append(total_sentiment)
sentiment_score.append(total_sentiment_score)
scores_df = pd.DataFrame(sentiment)
scores_df_1 = pd.DataFrame(sentiment_score)
parsed_and_scored_news = parsed_and_scored_news.join(scores_df, rsuffix='_right')
parsed_and_scored_news = parsed_and_scored_news.join(scores_df_1, rsuffix='_right')
st = parsed_and_scored_news['0_right'].tolist()
count = -1
for item in parsed_and_scored_news['0']:
count = count+1
if item == 'NEGATIVE':
lst[count] = 0-lst[count]
scores_final = pd.DataFrame(lst)
parsed_and_scored_news = parsed_and_scored_news.join(scores_final, rsuffix='_final')