Introduction: Easy Machine Learning Using Astrology

About: professional astrologer

“A year spent in artificial intelligence is enough to make one believe in God.” —Alan Perlis

In the past decade, a computer science technique called machine learning has been turning the heads of professionals ranging from linguists to actuaries to medical doctors to rocket scientists -- and for good reason. This very 21st century technology is yielding results that were not dreamed of just a few decades before.

What is machine learning? I am a fan of Wikipedia, so here is its definition: a computer science field that “gives computer systems the ability to ‘learn’ (i.e., progressively improve performance on a specific task) with data, without being explicitly programmed.” It is thus a type of process for generating artificial intelligence.

“By far, the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.” —Eliezer Yudkowsky

The math can get hairy. Luckily for you, me, and 98% of the professionals who use machine learning tools, just in the past three years or so, the system itself no longer needs to be completely developed by the end user.

I would like to show you, whether an astrologer or not, how to use these machine learning tools for the betterment of all. I will be focusing here on the free tools at which I have found to be at present the easiest to use with the best looking output while still using decent algorithms.

Taking less than twenty minutes of your time, this article will show how to use these tools with an engaging astrological data set.

Here, we will consider the small, simple, compelling, and public, people-oriented data set of 27 contemporary celebrity self-defined alcoholics and 27 contemporary celebrities who have a reputation of completely eschewing alcohol. Fairly extreme celebrity examples of each type were typically used.

Of course, once you run through this quick Instructable, you can use data of any type of your choosing.

The birth data are all from Astro-databank. See the appendix to this article for full birth details.

While I have some concern about the precision of the Astro-databank birth minutes when taken as a whole, the astrological data we will be using do not include the ascendant, and so perhaps a 1 to 60 minute approximation of birth minute in Astro-databank could be said to not impact too much the features of the very basic birth astronomy data in use.

These features are Sun degree, Moon degree, North Lunar Node degree, South Lunar Node degree, Mercury degree, Venus degree, Mars degree, Jupiter degree, Saturn degree, Uranus degree, Neptune degree, Pluto degree, Mercury retrograde true or false, Venus retrograde true or false, Mars retrograde true or false, Jupiter retrograde true or false, Saturn retrograde true or false, Uranus retrograde true or false, Neptune retrograde true or false, Pluto retrograde true or false.

All degrees are Tropical placements out of 360 degrees for that celebrity’s birth, computed via right ascensions using astronomy software.

When I ran the procedure but with Sidereal degrees (using Lahiri ayanamsha), the results were also positive, but not as clean and clear, and so for this presentation, we are using Tropical degrees.

Let’s do some AI.

Step 1: Create a New Project.

Set up a free account at Navigate to your dashboard. An introductory project will be shown, titled “BigML Intro Project”. That can be fun to explore, but for now, create a new project.

Step 2: Name Your Project.

I named mine “Celebrity Alcohol Use”. You do not need to fill out the tags or description, but you can if you wish.

Step 3: Download Astrology Data to Your Computer.

Download the celebrity alcohol data set.

Step 4: Upload Astrology Data to Your Project.

In your browser BigML page, click on the first icon to the left that says “Sources” when you mouse over it.

Step 5: Create a Source.

Now click on the icon to the far right that says upon mouseover “Create a source from a local file,” and upload your new data set.

Step 6: Create a Dataset Within From Your Source.

It will take about 10 seconds for the option to appear, but then you can choose the drop down next to the file name and choose “1-click dataset”.

Step 7: Open Your New Dataset.

Next, click on "Datasets" at the top gray bar.

On the next screen, mouse over "Supervised" in the gray bar at top and drop down and select "OPTIML". A faint box with a "1" will appear six columns to the right of your file name. Click on it.

You will see the third image as your screen. The process will take a good while to complete, so now is a great time to have a nice cup of tea.

Step 8: Look at Your Models.

After about 15 minutes, you will see this screen. These are the best performing models, different ways the computer has of understanding the data.

To see what the top model is based on, click on "boosted trees, 1235-node...." to open in a new tab. Alternatively, you can also see the role of the various mutually interacting astronomy factors by going here.

To see the top model's success, go back to the first tab in your browser and click on the icon to the right of the top green line, the one that says "View evaluation" when you mouse over it.

Now, you get the most important page, the final results, as seen in the second image.

This is a very successful hands-free model. (Click on second image to enlarge it).

What it says is that 4/5 or 80.0% of the alcoholics in the sample test group were predicted accurately, and just as importantly, 5/6 or 83.3% of the non-alcoholics were also predicted accurately.

There is a rich body of information online to explain the meaning of each of the other numbers.

Step 9: Conclusion

Besides, some other platforms of artificial intelligence that are free or have reasonable trials include Weka, Google TensorFlow, Amazon Web Services, Mathematica, and Rapid Miner. Each has its strengths. There are others even now, and there are bound to be more of such commercial applications of machine learning appearing every month for the foreseeable future.

If you are so inclined, you can even explore the options in a successful BigML model to use it for evaluations of friends and acquaintances or whomever. I would caution you, however, to think greatly of whether this model is ready for prime time. Is it truly capable of being a diagnostic tool? I absolutely think not.

First of all, I would not feel comfortable in using it in any way until the same results of high dual accuracy arrive but by using tens of thousands of test examples.

Second of all, it needs to be said: diagnostics is legally still in the domain of the physician.

Despite this rudimentary example, now may be the time nonetheless for our field to start having conversations about the ethical use of this technology, just as these types of conversations are happening in other fields.

“With artificial intelligence, we are summoning the demon.” —Elon Musk

Even with great results and the tantalizing prospect of re-establishing astrology as a biomarker like any other, let us not lose our spirit.

Still, I would say: go out there and explore this revolutionary technology. Use your own data, have fun, and don’t forget to report back to us your feelings as well as your findings.

Step 10: Appendix of Birth Charts Used in Spreadsheet

Row # Name Rodden Rating Birth Date Birth Time Birth Place Public History of Alcohol Use

2 Stephen King A Sep 21, 1947 1:30 am Portland, ME,USA True

3 Robin Williams AA Jul 21, 1951 13:34 pm Chicago, IL, USA True

4 Ben Affleck AA Aug 15, 1972 2:53 am Berkeley, CA, USA True

5 Michael J. Fox A Jul 9, 1961 0:15 am Edmonton, Alb., Canada True

6 Jamie Lee Curtis AA Nov 22, 1958 8:37 am Los Angeles, CA, USA True

7 Diana Ross AA Mar 26, 1944 23:46 pm Detroit,MI, USA True

8 Mel Gibson B Jan 3, 1956 16:45 pm Peekskill, NY USA True

9 Johnny Depp AA Jun 9, 1963 8:44 am Owensboro, KY, USA True

10 Billie Holiday AA Apr 7, 1947 2:30 am Philadelphia, PA, USA True

11 Anthony Hopkins A Dec 31, 1937 9:15 am Port Talbot, Wales, UK True

12 Robert Downey, Jr. A Apr 4, 1965 13:10 pm New York, NY True

13 Nick Nolte AA Feb 8, 1941 10:40 am Omaha, NE, USA True

14 Melanie Griffith A Aug 9, 1957 23:49 pm New York, NY True

15 Ewan McGregor AA Mar 31, 1971 20:10 pm Perth, Scotland, UK True

16 Betty Ford AA Apr 8, 1918 15:45 pm Chicago, IL, USA True

17 Mackenzie Phillips A Nov 10, 1959 21:35 pm Alexandria, VA, USA True

18 Ernest Hemingway AA Jul 21, 1899 8:00 am Oak Park, IL, USA True

19 Buzz Aldrin AA Jan 20, 1930 14:17 pm Glen Ridge, NJ, USA True

20 Elizabeth Taylor AA Feb 27, 1932 2:30 am London, UK True

21 Edie Sedgwick AA Apr 20, 1943 9:47 am Santa Barbara, CA, USA True

22 Amy Winehouse A Sep 14, 1983 22:25 pm Enfield, UK True

23 Matthew Perry AA Aug 19,1969 4:47 am North Adams, MA, USA True

24 Russell Brand B Jun 4, 1975 0:01 am Grays, UK True

25 Macklemore A Jun 19, 1983 7:45 am Seattle, WA, USA True

26 Judy Garland AA Jun 10, 1922 6:00 am Grand Rapids, MN True

27 Leon Smet AA Oct 27, 1923 11:00 am Schaerbeek, Belgium True

28 Alcoholic 5746 AA Oct 27, 1923 7:00 am Spokane, WA, USA True

29 Blake Lively AA Aug 25, 1987 5:07 am Tarzana, CA, USA False

30 Kim Kardashian AA Oct 21, 1987 10:46 am Los Angeles, CA, USA False

31 Tyra Banks AA Dec 4, 1973 19:17 pm Inglewood, CA, USA False

32 Gwyneth Paltrow AA Sep 27, 1972 17:25 pm Los Angeles, CA, USA False

33 Mindy Kaling AA Jun 24, 1979 4:10 am Waltham, MA, USA False

34 Connie Britton AA Mar 6, 1967 15:49 pm Boston, MA, USA False

35 John Legend AA Dec 28, 1978 8:25 am Springfield, OH, USA False

36 John Krasinski AA Oct 20, 1979 20:44 pm Brighton, MA, USA False

37 Jake Gyllenhaal AA Dec 19, 1980 20:08 pm Los Angeles, CA, USA False

38 Meryl Streep AA Jun 22, 1949 8:05 am Summit, NJ, USA False

39 Sigourney Weaver AA Oct 8, 1949 18:15 pm New York, NY, USA False

40 Vanessa Bayer AA Nov 14, 1981 6:50 am Cleveland, OH, USA False

41 Kathryn Bigelow AA Nov 27, 1951 7:19 am San Mateo, CA, USA False

42 Jodie Foster AA Nov 19, 1962 8:14 am Los Angeles, CA, USA False

43 LeBron James AA Dec 30, 1984 16:04 pm Akron, OH, USA False

44 George Clooney AA May 6, 1961 2:58 am Lexington, KY, USA False

45 Conan O’Brien AA Apr 18, 1963 13:48 pm Boston, MA, USA False

46 Jennifer Lawrence AA Aug 15, 1990 15:20 pm Louisville, KY, USA False

47 Tom Hanks AA Jul 9, 1956 11:17 am Concord, CA,USA False

48 Jennifer Anniston AA Feb 11, 1969 22:22 pm Los Angeles, CA, USA False

49 Jimmy Carter AA Oct 1, 1924 7:00 am Plains, GA, USA False

50 Prince AA Jun 7, 1958 18:17 pm Minneapolis, MN, USA False

51 Salvador Dali AA May 11, 1904 8:45 am Figueras, Spain False

52 Barbara Streisand AA Apr 24, 1942 5:08 am New York, NY, USA False

53 Neil Diamond AA Jan 24, 19411 23:04 pm New York, NY, USA False

54 Emma Watson AA Apr 15, 1990 18:00 pm Paris, France False

55 Agatha Christie AA Sep 15, 1890 14:14 pm Torquay, UK False