Skip to content

Business Events Knowledge Model

Overview

The Business Events Knowledge Model (display name: Business Events EN v#) extracts entities related to business operations and events for building a company financial profile. It also classifies documents dealing with corporate events.

The purpose of this model is to help drawing an insightful overview of a company's main business and financial actions.

The model performs both categorization and extraction.

Categorization

The topic of financial facts is broad, ranging from personal finance to financial institution indices.
The model covers the main areas of interest in analyzing a company's financial performance and health, particularly from the perspective of a potential shareholder.

Tip

In addition to ratios and indices, Nasdaq suggests considering a company's business model and management changes.

The taxonomy is composed of six main thematic clusters, under the root node EVENTS_TAXONOMY, as follows:

10000 EVENTS_TAXONOMY
    11000 Management and Ownership
        11100 Board of Directors and Management
            11110 New Appointment
            11120 Resignation
            11130 Meeting
        11200 Newly Public Companies
        11300 Dividends
        11400 Mergers and Acquisitions
        11500 New Shareholders
    12000 Globalization
        12100 Offshoring
        12200 Greenfield FDIs
        12300 Brownfield FDIs
    13000 Business Operations
        13100 Partnership and Trade Agreements
        13200 New Products or Services
        13300 Stock Offering
        13400 Patenting, Copyright and Trademark
        13500 Sales Increase
        13600 Research and Development
        13700 Financial Statements
            13710 Net Income
            13720 Net Profits
        13800 Business Models
        13900 New Openings
    14000 Corporate Profile
        14100 Social Popularity
        14200 Awards
        14300 Legal Issues
        14400 Negative Opinions
        14500 Charity and Social Initiatives
    15000 Financial Instability
        15100 Downsizing
        15200 Layoffs
        15300 Losses and Sales Decrease
        15400 Bankruptcy
    16000 Stock Market Trends

where:

  • Management and Ownership covers events that involve board of directors, management and ownership changes, including mergers, acquisitions and newly public companies.
  • Globalization refers to FDIs and outsourcing.
  • Business Operations includes all of those transversal operations that aim at creating value from physical or intangible assets.
  • Corporate Profile includes the corporate image and reputation perceived by the public, as well as financial advisors.
  • Financial Instability covers critical situations, like downsizing or layoffs.
  • Stock Market Trends is used to categorize documents that report stock market news.

Extraction groups and classes

Extraction has the goal to detect entities linked to the main financial facts, like CEO statements or stock trends. It is structured with the following groups and their related classes.

MANAGEMENT_DIRECTION

The classes of the MANAGEMENT_DIRECTION group correspond to information about managers and directors of a company.
They are:

Class Description
quote Any quote by a manager/director
resignation Job role left by a manager/director
new_appointment New role for which a manager/director was hired
role Job role filled by a manager/director within the organization.
firm Company name
who Manager/director name

STOCK_SWINGS

The classes of the STOCK_SWINGS group correspond to names and tickers of the major traded stocks, as well as the related market trends—also in percentage—mentioned in financial news and articles.
They are:

Class Description
bearish Downward trend
bullish Upward trend
stock Stock name

Classes are extracted as follows:

  • When there are no trends associated with the stock mentioned in the text, only the stock class will be triggered.
  • When there is a bullish or bearish trend related to a specific stock, with no information about price increase/decrease, the bearish or bullish classes produce an extraction containing the name of the stock or just the lemma stock when the text mentions the company but not the stock name.
  • If the reported trend includes a percentage, the stock class produces an extraction with the name of the related stock—or just the lemma stock if the text mentions the company but not the stock name—while the class bearish or bullish classes show the related percentage.

ENTITIES

The classes of the ENTITIES group correspond to the companies in a documents and possible quotes about the company and the mass media name where the quotes appeared.
They are:

Class Description
company Name of the company
media Media name
statement Text of the quote

Output structure

The model output has the same structure as any other model and is affected by the functional properties of the workflow block.
The peculiar parts of the output are the result of categorization, i.e. the categories array, and the result of information extraction, i.e. the extractions array.

Example

Considering the sentence:

L'Oréal is supposed to "change its business model", The Financial Times says.

The categorization is:

Category ID Label
13800 Business Models

The extraction is:

Group Class Value
ENTITIES company L'Oréal
ENTITIES media The Financial Times
ENTITIES statement change its business model

In this model's output, the template key corresponds to the concept of group and template fields correspond to classes.

"categories": [
    {
        "frequency": 0,
        "hierarchy": [
            "Business Operations",
            "Business Models"
        ],
        "id": "13800",
        "label": "Business Models",
        "namespace": "events_en",
        "positions": [
            {
                "end": 7,
                "start": 0
            },
            {
                "end": 49,
                "start": 35
            }
        ],
        "score": 20,
        "winner": true
    }
],
"content": "L'Oréal is supposed to \"change its business model\", The Financial Times says\".\n",
"entities": [
    {
        "lemma": "The Financial Times",
        "positions": [
            {
                "end": 71,
                "start": 52
            }
        ],
        "syncon": 112270,
        "type": "MMD"
    },
    {
        "lemma": "L'Oréal",
        "positions": [
            {
                "end": 7,
                "start": 0
            }
        ],
        "syncon": 289289,
        "type": "COM"
    }
],
"extractions": [
    {
        "fields": [
            {
                "name": "company",
                "positions": [
                    {
                        "end": 7,
                        "start": 0
                    }
                ],
                "value": "L'Oréal"
            },
            {
                "name": "media",
                "positions": [
                    {
                        "end": 71,
                        "start": 52
                    }
                ],
                "value": "The Financial Times"
            },
            {
                "name": "statement",
                "positions": [
                    {
                        "end": 50,
                        "start": 23
                    }
                ],
                "value": "\"change its business model\""
            }
        ],
        "namespace": "events_en",
        "template": "ENTITIES"
    }
]