OK! There is a Caveat. NOT in Suicide ‘Prevention’ Per Se, But Rather ‘Imminent Suicide Detection’ aka ‘Imminent Suicidal Tendency Classification’
One person dies by suicide every 40 seconds: UN health agency report
One person dies by suicide every 40 seconds: new UN health agency report
Speaking in Geneva ahead of World Suicide Prevention Day, recognized on 10 September, Dr. Tedros Adhanom Ghebreyesus…
3 Times as many Men as Women…
~800,000 people commit suicide every year.
Low and Middle Income Nations Accounted for ~75% Cases All Around The World.
Most Common Methods
- Self Poisoning
- Fire Arms
Only 28 Countries Worldwide Have National Suicide Prevention Strategies.
Most Common Age Group 70+
Suicide is the second most leading cause of Death in Young People Aged 15–29.
10th September is The World Suicide Prevention Day.
First and Foremost, we needed a way to categorize and classify Human Emotions!!!
Then we needed to order them on a scale of Mild, Moderate and Strong Emotions. So we finally arrived at this…
We got 10 categories of Emotions, within which we were concerned with the ‘Strong’ Emotions i.e. Row 1.
Of which we eliminated ‘Happy’ & ‘Caring’ Emotions (Category and Items)
The Algorithm — Hierarchical 2 Step Classification
Phase — 1 — Core Emotion Classification
First we classify — the text from (Chat/Posts/Emails…) into Basic Emotions of 8 ‘Strong’ Emotion Categories.
This can be done using any Basic NLP Algorithm similar to Corpus Based Sentiment Analysis Algorithms for the sake of simplicity. To various degrees of accuracy.
Phase — 2 — Universal Classification of Aggregate Emotions Indicating Suicidal Tendency(s)
This is where we use our Universal Classification Algorithm.
We combine 100+ Dimensions of Emotions on a scale of 0–100 into One Single Binary True/False Classification of Imminent Suicidal Tendency.
And the way this works is that we make up for any loss of accuracy in phase-1 and yet achieve ~100% accuracy in the Classification.
We needed an Intervention Strategy (From our Empirical Research)
Long Term Strategy
It is known that Restricting Access to Means of Suicide is Reasonably Effective.
The Moment Strategy
If we can prevent Suicide on the Spur of The Moment, at the Peak of Suicidal Intentions. The Suicidal Tendency Mostly Fades Away to Occasional Mild Tendencies.
We also know that The World Will Probably Get Destroyed on The Spur of The Moment. Because we know The Longer The Time/Delay, Exponentially More Rationality Will Prevail.
As Suicide Gets Enacted, it crosses the point of no return. And the act becomes irreversible. But before Death the Patient shows Regret and Hopes and Prays to Let God Reverse his Decision, he wants to live. But it becomes too late.
We also needed to Reverse Infer Causes
- Situational, versus
In an internal, or dispositional, attribution, people infer that an event or a person’s behavior is due to personal factors such as traits, abilities, or feelings. In an external, or situational, attribution, people infer that a person’s behavior is due to situational factors.
We Then Needed To Find a Fix For The Root Causes
Most Common Situational Causes are Love, Money, Reputation/Image, Loss, Bullying, Isolation/Boycott, Social/Peer Pressure etc.
*** Fixing Situations almost take a lifetime because No One Can Single Handedly Change The World.
Most Common Dispositional Causes are Inability, Inadequacy, Deficiency, Disability, Failures etc.
Situational Causes Normally Take Between 5–20 years to Fix/Get-Fixed.
Dispositional Causes Can Be ‘Stably’ Fixed With Minimal Chances of Relapse in 2–3 years.
We can Simply Passively/Non-Intrusively Observe and Identify/Classify Emotions.
We are now able to classify (Classification in Machine Learning) with ~100% Accuracy, Human Emotions. Which can then activate the Intervention(s).
Here are some related announcements and related breakthroughs from us earlier.
Millennium Breakthrough — How we created Universal Learning and came a step closer to Artificial…
What we have today are systems that are trained on specific input/output data, lots of it, to learn a non-linear…
How We Achieved ~100% Accuracy In “Cancer Detection” Besides Theoretical Best Accuracy in Credit…
Model Complete (Logic/Turing Complete)
Millennium Breakthrough — Automatski Announces The General Availability of Universal…
What is Classification?
This is our website http://automatski.com