def test_extract_txt(reload=False):
if reload:
best_vector = "wordfreq"
best_model = 1 # linearLogistic
save_model(best_vector, best_model)
else:
od.loadStopwords()
od.loadEmotionwords()
od.loadWords(od.stopList)
od.loadDocument(od.stopList)
predictor = Predictor()
predictor.load_model()
predictor.set_mode(mode="wordfreq")
# 读取txt文件
path = os.path.join("reports")
file_names = os.listdir(path)
for file in file_names:
if file.lower().count('.txt') == 1:
result = []
with open(os.path.join(path, file), 'r', encoding="utf-8") as f:
contents = f.read()
news_list = format_text(contents)
for new in news_list:
predictor.set_news(news=new)
predictor.trans_vec()
tag = predictor()
result.append((new, tag))
# 计算贸易摩擦指数
print(file)
total, neg_nums, percent = cal_neg_ratio(result)
print("贸易摩擦指数：")
print(percent)
write_result(file, total, neg_nums, percent)
def save_model(best_vector, best_model):
od.loadStopwords()
od.loadEmotionwords()
od.loadWords(od.stopList)
od.loadDocument(od.stopList)
xpath = os.path.join('result', 'vector', 'resultX.npz')
ypath = os.path.join('result', 'vector', 'resultY.npz')
resultX = np.load(xpath)
resultY = np.load(ypath)
new_x, new_y = od.twoTag(resultX[best_vector], resultY[best_vector])
model_saved = ml.naiveBayes(new_x, new_y)
path = os.path.join('model', 'wordfreq_naiveBayes.ml')
with open(path, 'wb') as f:
pickle.dump(model_saved, f)
print("Save over")
def save_model(best_vector, best_model):
"""
存储效果最好的模型
需要手动指明参数名字
:param best_vector: 最好的文本->词向量的方法
:param best_model: 最好的机器学习模型
:return: info
"""
od.loadStopwords()
od.loadEmotionwords()
od.loadWords(od.stopList)
od.loadDocument(od.stopList)
xpath = os.path.join('result', 'vector', 'resultX.npz')
ypath = os.path.join('result', 'vector', 'resultY.npz')
resultX = np.load(xpath)
resultY = np.load(ypath)
new_x, new_y = od.twoTag(resultX[best_vector], resultY[best_vector])
model_saved = ml.linearLogistic(new_x, new_y)
path = os.path.join('model', 'wordfreq_logistic.ml')
with open(path, 'wb') as f:
pickle.dump(model_saved, f)
print("Save over")
def test(reload=False):
if reload:
best_vector = "wordfreq"
best_model = 1 # linearLogistic
save_model(best_vector, best_model)
else:
od.loadStopwords()
od.loadEmotionwords()
od.loadWords(od.stopList)
od.loadDocument(od.stopList)
predictor = Predictor()
predictor.load_model()
predictor.set_mode(mode="wordfreq")
news = " 《经济通通讯社13日专讯》日股早市偏软,日经225指数报18312跌239点。 美元兑日圆疲软,新报108﹒78╱80。(tt) "
news = " 周二,恒生指数收报20356.24点,跌236.76点,跌幅1.15%;国企指数收报10596.91点,跌148点,跌幅1.38%;大市成交492.76亿港元。美国3月非农就业数据表现疲弱,拖累隔夜欧美股市全线受压。中国3月份CPI同比增长3.6%,令货币政策在短期内放宽预期降低。港股早盘随外围低开两百多点,但是A股在汇金增持内银股刺激下探底回升,对港股起到支持,之后恒指于低位维持窄幅震荡整理,最终跌逾1%。银行股全线走软。四大内银股方面,工商银行跌0.4%,中国银行跌0.64%,建设银行跌0.67%,农业银行跌1.2%;国际金融股方面,汇丰控股跌1.75%,渣打集团跌1.89%。美国就业市场增长放缓及内地通胀反弹,投资者对经济信心下降。中国央行短期内下调存准机会大减,从而利淡大市气氛。预计港股本周将继续在20200至20700点之间震荡。 " # 待转化的文本
predictor.set_news(news=news)
predictor.trans_vec()
tag = predictor()
print("算出来的和是", sum(predictor._vec[0]))
print("打标的结果是：", tag)