From 29347ce7fbd55a47c3271f1afc33e7feec934da6 Mon Sep 17 00:00:00 2001 From: Philipp Horstenkamp Date: Mon, 1 May 2023 12:20:56 +0200 Subject: [PATCH] Added a text in research central. --- .../Sentiment Analysis/FinBert.ipynb | 457 +++++++----------- documentations/research/resarch-central.md | 12 + 2 files changed, 180 insertions(+), 289 deletions(-) diff --git a/Jupyter/AI-models/Sentiment Analysis/FinBert.ipynb b/Jupyter/AI-models/Sentiment Analysis/FinBert.ipynb index 27b5429..a5e213f 100644 --- a/Jupyter/AI-models/Sentiment Analysis/FinBert.ipynb +++ b/Jupyter/AI-models/Sentiment Analysis/FinBert.ipynb @@ -6,19 +6,41 @@ "source": [ "# FinBert\n", "\n", + "FinBert is a sentiment Analysis AI for Financial text.\n", + "Since we want to evaluate news article this is a necessary feature to analyse those texts.\n", + "In this document a first use of this tool will be shown.\n", + "Some texts will be analysed. Especially the analysis of german texts will be tried.\n", + "\n", "## Sources\n", "\n", "[HugginFace](https://huggingface.co/ProsusAI/finbert)\n", "[Tutorial](https://medium.com/codex/stocks-news-sentiment-analysis-with-deep-learning-transformers-and-machine-learning-cdcdb827fc06)" ] }, + { + "cell_type": "markdown", + "source": [ + "## Libraries\n", + "\n", + "* transformers\n", + "* tqdm\n", + "* pandas\n", + "* numpy\n", + "* torch\n", + "* torchvision\n", + "* torchaudio" + ], + "metadata": { + "collapsed": false + } + }, { "cell_type": "code", - "execution_count": 13, + "execution_count": 1, "metadata": { "ExecuteTime": { - "end_time": "2023-04-30T21:54:44.056694Z", - "start_time": "2023-04-30T21:53:45.027971Z" + "start_time": "2023-05-01T12:07:42.348877Z", + "end_time": "2023-05-01T12:07:46.944359Z" }, "collapsed": false, "jupyter": { @@ -27,12 +49,47 @@ "tags": [] }, "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Requirement already satisfied: transformers in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (4.28.1)\n", + "Requirement already satisfied: tqdm in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (4.65.0)\n", + "Requirement already satisfied: pandas in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (2.0.1)\n", + "Requirement already satisfied: numpy in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (1.24.3)\n", + "Requirement already satisfied: torch in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (2.0.0)\n", + "Requirement already satisfied: torchvision in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (0.15.1)\n", + "Requirement already satisfied: torchaudio in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (2.0.1)\n", + "Requirement already satisfied: filelock in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from transformers) (3.8.0)\n", + "Requirement already satisfied: huggingface-hub<1.0,>=0.11.0 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from transformers) (0.14.1)\n", + "Requirement already satisfied: packaging>=20.0 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from transformers) (23.1)\n", + "Requirement already satisfied: pyyaml>=5.1 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from transformers) (6.0)\n", + "Requirement already satisfied: regex!=2019.12.17 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from transformers) (2023.3.23)\n", + "Requirement already satisfied: requests in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from transformers) (2.28.1)\n", + "Requirement already satisfied: tokenizers!=0.11.3,<0.14,>=0.11.1 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from transformers) (0.13.3)\n", + "Requirement already satisfied: colorama in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from tqdm) (0.4.6)\n", + "Requirement already satisfied: python-dateutil>=2.8.2 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from pandas) (2.8.2)\n", + "Requirement already satisfied: pytz>=2020.1 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from pandas) (2022.7)\n", + "Requirement already satisfied: tzdata>=2022.1 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from pandas) (2023.3)\n", + "Requirement already satisfied: typing-extensions in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from torch) (4.5.0)\n", + "Requirement already satisfied: sympy in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from torch) (1.11.1)\n", + "Requirement already satisfied: networkx in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from torch) (3.1)\n", + "Requirement already satisfied: jinja2 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from torch) (3.1.2)\n", + "Requirement already satisfied: pillow!=8.3.*,>=5.3.0 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from torchvision) (9.4.0)\n", + "Requirement already satisfied: fsspec in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from huggingface-hub<1.0,>=0.11.0->transformers) (2023.4.0)\n", + "Requirement already satisfied: six>=1.5 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from python-dateutil>=2.8.2->pandas) (1.16.0)\n", + "Requirement already satisfied: MarkupSafe>=2.0 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from jinja2->torch) (2.1.2)\n", + "Requirement already satisfied: charset-normalizer<3,>=2 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from requests->transformers) (2.1.1)\n", + "Requirement already satisfied: idna<4,>=2.5 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from requests->transformers) (3.4)\n", + "Requirement already satisfied: urllib3<1.27,>=1.21.1 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from requests->transformers) (1.26.12)\n", + "Requirement already satisfied: certifi>=2017.4.17 in c:\\users\\phhor\\appdata\\roaming\\python\\python311\\site-packages (from requests->transformers) (2022.9.24)\n", + "Requirement already satisfied: mpmath>=0.19 in c:\\users\\phhor\\pycharmprojects\\aki_prj23_transparenzregister\\venv\\lib\\site-packages (from sympy->torch) (1.3.0)\n" + ] + }, { "name": "stderr", "output_type": "stream", "text": [ - "ERROR: To modify pip, please run the following command:\n", - "C:\\Users\\phhor\\PycharmProjects\\aki_prj23_transparenzregister\\venv\\Scripts\\python.exe -m pip install transformers tqdm pandas numpy torch torchvision torchaudio pip -Uq\n", "\n", "[notice] A new release of pip is available: 23.0.1 -> 23.1.2\n", "[notice] To update, run: python.exe -m pip install --upgrade pip\n" @@ -40,22 +97,46 @@ } ], "source": [ - "!pip install transformers tqdm pandas numpy torch torchvision torchaudio pip -Uq" + "!pip install transformers tqdm pandas numpy torch torchvision torchaudio -U" ] }, + { + "cell_type": "markdown", + "source": [ + "### Importing and creation of models and tokenizer" + ], + "metadata": { + "collapsed": false + } + }, { "cell_type": "code", - "execution_count": 16, + "execution_count": 2, "metadata": { "collapsed": false, "jupyter": { "outputs_hidden": false }, - "tags": [] + "tags": [], + "ExecuteTime": { + "start_time": "2023-05-01T12:07:46.944359Z", + "end_time": "2023-05-01T12:07:51.049695Z" + } }, - "outputs": [], + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "C:\\Users\\phhor\\PycharmProjects\\aki_prj23_transparenzregister\\venv\\Lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n", + " from .autonotebook import tqdm as notebook_tqdm\n" + ] + } + ], "source": [ "import pandas as pd\n", + "import torch\n", + "\n", "from transformers import AutoTokenizer, AutoModelForSequenceClassification\n", "\n", "# create a tokenizer object\n", @@ -65,54 +146,26 @@ "model = AutoModelForSequenceClassification.from_pretrained(\"ProsusAI/finbert\")" ] }, + { + "cell_type": "markdown", + "source": [ + "### Analyze a single sentiment" + ], + "metadata": { + "collapsed": false + } + }, { "cell_type": "code", - "execution_count": 12, + "execution_count": 3, "metadata": { "collapsed": false, "jupyter": { "outputs_hidden": false }, - "tags": [] - }, - "outputs": [ - { - "data": { - "text/plain": [ - "tensor([[0.0535, 0.0279, 0.9185]], grad_fn=)" - ] - }, - "execution_count": 12, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "# A headline to be used as input\n", - "import torch\n", - "\n", - "headline = \"Microsoft fails to hit profit expectations\"\n", - "headline2 = (\n", - " \"Am Aktienmarkt überwieg weiter die Zuversicht, wie der Kursverlauf des DAX zeigt.\"\n", - ")\n", - "\n", - "# Pre-process input phrase\n", - "input_tokens = tokenizer(headline2, padding=True, truncation=True, return_tensors=\"pt\")\n", - "# Run inference on the tokenized phrase\n", - "output = model(**input_tokens)\n", - "\n", - "# Pass model output logits through a softmax layer.\n", - "sentim_scores = torch.nn.functional.softmax(output.logits, dim=-1)\n", - "sentim_scores" - ] - }, - { - "cell_type": "code", - "execution_count": 56, - "metadata": { - "collapsed": false, - "jupyter": { - "outputs_hidden": false + "ExecuteTime": { + "start_time": "2023-05-01T12:07:51.050690Z", + "end_time": "2023-05-01T12:07:51.129723Z" } }, "outputs": [ @@ -125,20 +178,15 @@ }, { "data": { - "text/plain": [ - "+ 0.034084\n", - "0 0.932933\n", - "- 0.032982\n", - "dtype: float32" - ] + "text/plain": "+ 0.034084\n0 0.932933\n- 0.032982\ndtype: float32" }, - "execution_count": 56, + "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "def analyze_sentiment(text: str):\n", + "def analyze_sentiment(text: str) -> pd.Series:\n", " print(text)\n", " input_tokens = tokenizer(text, padding=True, truncation=True, return_tensors=\"pt\")\n", " output = model(**input_tokens)\n", @@ -148,106 +196,37 @@ " )\n", "\n", "\n", + "headline = \"Microsoft fails to hit profit expectations\"\n", "tf = analyze_sentiment(headline)\n", "tf" ] }, { - "cell_type": "code", - "execution_count": 80, + "cell_type": "markdown", + "source": [ + "### Creating test data" + ], "metadata": { - "tags": [] + "collapsed": false + } + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": { + "tags": [], + "ExecuteTime": { + "start_time": "2023-05-01T12:07:51.130725Z", + "end_time": "2023-05-01T12:07:51.193306Z" + } }, "outputs": [ { "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
textlan
0Microsoft fails to hit profit expectationsen
1Am Aktienmarkt überwieg weiter die Zuversicht,...de
2Stocks rallied and the British pound gained.en
3Meyer Burger bedient ab sofort australischen M...de
4Meyer Burger enters Australian market and exhi...en
5J&T Express Vietnam hilft lokalen Handwerksdör...en
67 Experten empfehlen die Aktie zum Kauf, 1 Exp...de
7Microsoft aktie fällt.de
8Microsoft aktie steigt.de
\n", - "
" - ], - "text/plain": [ - " text lan\n", - "0 Microsoft fails to hit profit expectations en\n", - "1 Am Aktienmarkt überwieg weiter die Zuversicht,... de\n", - "2 Stocks rallied and the British pound gained. en\n", - "3 Meyer Burger bedient ab sofort australischen M... de\n", - "4 Meyer Burger enters Australian market and exhi... en\n", - "5 J&T Express Vietnam hilft lokalen Handwerksdör... en\n", - "6 7 Experten empfehlen die Aktie zum Kauf, 1 Exp... de\n", - "7 Microsoft aktie fällt. de\n", - "8 Microsoft aktie steigt. de" - ] + "text/plain": " text lan\n0 Microsoft fails to hit profit expectations en\n1 Am Aktienmarkt überwieg weiter die Zuversicht,... de\n2 Stocks rallied and the British pound gained. en\n3 Meyer Burger bedient ab sofort australischen M... de\n4 Meyer Burger enters Australian market and exhi... en\n5 J&T Express Vietnam hilft lokalen Handwerksdör... en\n6 7 Experten empfehlen die Aktie zum Kauf, 1 Exp... de\n7 Microsoft aktie fällt. de\n8 Microsoft aktie steigt. de", + "text/html": "
\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
textlan
0Microsoft fails to hit profit expectationsen
1Am Aktienmarkt überwieg weiter die Zuversicht,...de
2Stocks rallied and the British pound gained.en
3Meyer Burger bedient ab sofort australischen M...de
4Meyer Burger enters Australian market and exhi...en
5J&T Express Vietnam hilft lokalen Handwerksdör...en
67 Experten empfehlen die Aktie zum Kauf, 1 Exp...de
7Microsoft aktie fällt.de
8Microsoft aktie steigt.de
\n
" }, - "execution_count": 80, + "execution_count": 4, "metadata": {}, "output_type": "execute_result" } @@ -284,13 +263,39 @@ "text_df" ] }, + { + "cell_type": "markdown", + "source": [], + "metadata": { + "collapsed": false + } + }, { "cell_type": "code", - "execution_count": 81, + "execution_count": 6, + "outputs": [], + "source": [ + "### Analyze multiple Sentiments" + ], + "metadata": { + "collapsed": false, + "ExecuteTime": { + "start_time": "2023-05-01T12:10:36.719000Z", + "end_time": "2023-05-01T12:10:36.725700Z" + } + } + }, + { + "cell_type": "code", + "execution_count": 5, "metadata": { "collapsed": false, "jupyter": { "outputs_hidden": false + }, + "ExecuteTime": { + "start_time": "2023-05-01T12:07:51.151294Z", + "end_time": "2023-05-01T12:07:51.992442Z" } }, "outputs": [ @@ -311,142 +316,17 @@ }, { "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
textlan+0-
0Microsoft fails to hit profit expectationsen0.0340840.9329330.032982
1Am Aktienmarkt überwieg weiter die Zuversicht,...de0.0535280.0279500.918522
2Stocks rallied and the British pound gained.en0.8983610.0344740.067165
3Meyer Burger bedient ab sofort australischen M...de0.1165970.0127900.870613
4Meyer Burger enters Australian market and exhi...en0.1875270.0088460.803627
5J&T Express Vietnam hilft lokalen Handwerksdör...en0.0662770.0206080.913115
67 Experten empfehlen die Aktie zum Kauf, 1 Exp...de0.0503460.0220040.927650
7Microsoft aktie fällt.de0.0660610.0164400.917498
8Microsoft aktie steigt.de0.0414490.0184710.940080
\n", - "
" - ], - "text/plain": [ - " text lan + 0 \n", - "0 Microsoft fails to hit profit expectations en 0.034084 0.932933 \\\n", - "1 Am Aktienmarkt überwieg weiter die Zuversicht,... de 0.053528 0.027950 \n", - "2 Stocks rallied and the British pound gained. en 0.898361 0.034474 \n", - "3 Meyer Burger bedient ab sofort australischen M... de 0.116597 0.012790 \n", - "4 Meyer Burger enters Australian market and exhi... en 0.187527 0.008846 \n", - "5 J&T Express Vietnam hilft lokalen Handwerksdör... en 0.066277 0.020608 \n", - "6 7 Experten empfehlen die Aktie zum Kauf, 1 Exp... de 0.050346 0.022004 \n", - "7 Microsoft aktie fällt. de 0.066061 0.016440 \n", - "8 Microsoft aktie steigt. de 0.041449 0.018471 \n", - "\n", - " - \n", - "0 0.032982 \n", - "1 0.918522 \n", - "2 0.067165 \n", - "3 0.870613 \n", - "4 0.803627 \n", - "5 0.913115 \n", - "6 0.927650 \n", - "7 0.917498 \n", - "8 0.940080 " - ] + "text/plain": " text lan + 0 \n0 Microsoft fails to hit profit expectations en 0.034084 0.932933 \\\n1 Am Aktienmarkt überwieg weiter die Zuversicht,... de 0.053528 0.027950 \n2 Stocks rallied and the British pound gained. en 0.898361 0.034474 \n3 Meyer Burger bedient ab sofort australischen M... de 0.116597 0.012790 \n4 Meyer Burger enters Australian market and exhi... en 0.187527 0.008846 \n5 J&T Express Vietnam hilft lokalen Handwerksdör... en 0.066277 0.020608 \n6 7 Experten empfehlen die Aktie zum Kauf, 1 Exp... de 0.050346 0.022004 \n7 Microsoft aktie fällt. de 0.066061 0.016440 \n8 Microsoft aktie steigt. de 0.041449 0.018471 \n\n - \n0 0.032982 \n1 0.918522 \n2 0.067165 \n3 0.870613 \n4 0.803627 \n5 0.913115 \n6 0.927650 \n7 0.917498 \n8 0.940080 ", + "text/html": "
\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
textlan+0-
0Microsoft fails to hit profit expectationsen0.0340840.9329330.032982
1Am Aktienmarkt überwieg weiter die Zuversicht,...de0.0535280.0279500.918522
2Stocks rallied and the British pound gained.en0.8983610.0344740.067165
3Meyer Burger bedient ab sofort australischen M...de0.1165970.0127900.870613
4Meyer Burger enters Australian market and exhi...en0.1875270.0088460.803627
5J&T Express Vietnam hilft lokalen Handwerksdör...en0.0662770.0206080.913115
67 Experten empfehlen die Aktie zum Kauf, 1 Exp...de0.0503460.0220040.927650
7Microsoft aktie fällt.de0.0660610.0164400.917498
8Microsoft aktie steigt.de0.0414490.0184710.940080
\n
" }, - "execution_count": 81, + "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "def analyse_sentiments(texts: pd.Series) -> pd.DataFrame:\n", + "def analyse_sentiments(texts: pd.DataFrame) -> pd.DataFrame:\n", " values = texts[\"text\"].apply(analyze_sentiment)\n", - " # print(values)\n", " texts[[\"+\", \"0\", \"-\"]] = values\n", " return texts\n", "\n", @@ -455,18 +335,17 @@ ] }, { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] + "cell_type": "markdown", + "source": [ + "## Conclusion\n", + "\n", + "The current form of this model can't be used for the german language.\n", + "It could be used if the text is translated beforehand. But it is questionable if that will work well.\n", + "Another way would be to retrain the same model with translated text from this models' data. But I do not believe this to be feasible." + ], + "metadata": { + "collapsed": false + } } ], "metadata": { diff --git a/documentations/research/resarch-central.md b/documentations/research/resarch-central.md index e69de29..14f1753 100644 --- a/documentations/research/resarch-central.md +++ b/documentations/research/resarch-central.md @@ -0,0 +1,12 @@ +# Research Central + + +## Sentiment Analysis + +### FinBert + +FinBert is a specialised sentiment Analysis for Financial Data. +Sadly it isn't a very good model, and it does not work at all for texts in german. + +Experiments can be found here: +* [FinBert Jupyter](../../Jupyter/AI-models/"Sentiment Analysis"/FinBert.ipynb)