S&P500 Market Stock Prediction With NYT Headlines
NLP with daily New York Times headlines
@kaggle.jonathanpoli_sp500_market_stock_prediction_with_nyt_headlines
NLP with daily New York Times headlines
@kaggle.jonathanpoli_sp500_market_stock_prediction_with_nyt_headlines
This dataset contains every headlines from the "World" section of the NYT for each day, concatenated in one column, ordered by popularity. The label represents if the S&P500 market is going to increase (1) or decrease (0) on the following day.
The aim is then to predict the market evolution based on the last news.
The metrics used for evaluating the model are the accuracy and the f1-score.
Extra-Data can be used.
Good luck and enjoy !
CREATE TABLE dataset_nytimes (
"date" TIMESTAMP,
"label" BIGINT,
"news" VARCHAR
);CREATE TABLE n_2008 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2009 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2010 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2011 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2012 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2013 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2014 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2015 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2016 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2017 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2018 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2019 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2020 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);CREATE TABLE n_2021 (
"unnamed_0" BIGINT -- Unnamed: 0,
"date" TIMESTAMP,
"news" VARCHAR
);Anyone who has the link will be able to view this.