Blog on GitHub Pages

What story UFO observational data is telling about the UFO phenomenon?

22 Dec 2019

 

1. Introduction 

1.1 Background

According to Wikipedia [1], UFO is the popular term for any aerial phenomenon that cannot immediately be identified. The term “UFO” was created in 1953 by the United States Air Force and in its initial definition, it was “any airborne object which by performance, aerodynamic characteristics, or unusual features, does not conform to any presently known aircraft or missile type, or which cannot be positively identified as a familiar object.” The Oxford English Dictionary defines a UFO as “An unidentified flying object; a ‘flying saucer’.” In widely known usage, the term UFO corresponds to claims of alien spacecraft.

Unexplained aerial observations have been described throughout history. For example, on April 14, 1561, residents of Nuremberg described the appearance of a large black triangular object. They described also hundreds of spheres, cylinders and other odd-shaped objects that moved erratically overhead. On January 25, 1878, the Denison Daily News published an article in which John Martin, a local farmer, had reported seeing a large, dark, circular object resembling a balloon flying “at wonderful speed.” It appeared to be about the size of a saucer, one of the first uses of the word “saucer” in association with a UFO. These are just two examples. Some of these observations were undeniably astronomical in nature: comets, bright meteors, one or more of the five planets that can be freely seen with the naked eye, planetary conjunctions (an apparent meeting or passing of two or more planets), or atmospheric optical phenomena such as parhelia (a phenomenon that consists of a bright spot to one or both sides of the Sun) and lenticular clouds (lentil shaped stationary clouds).

The majority of UFO observations are, according to studies, misidentified conventional objects or natural phenomena - like aircrafts, balloons, satellites, or meteors, bright stars, and planets. A small percentage are hoaxes and fewer than 10% of reported sightings remain unexplained after proper investigation, and therefore can be classified as unidentified in the strictest sense. However, the null hypothesis cannot be excluded for these reports in a sense that they are simply other phenomena that cannot be identified due to lack of respective information. Advocates of the extraterrestrial hypothesis suggest that these unexplained reports are of alien spacecraft. Almost no scientific papers about UFOs have been published in peer-reviewed journals. The general conclusion in the scientific community was that the phenomenon was not worthy of serious investigation except as a cultural artifact.

[1].https://en.wikipedia.org/wiki/Unidentified_flying_object

1.2 Problem

In this project we will try to reveal a story the UFO observational data is telling about the UFO phenomenon. When the above general conclusion was reached data science as a discipline did not exist. There are now publicly available datasets about UFO observations which can be freely used. If we combine these two ingredients, a free available UFO dataset and data science we can unlock insights about the UFO phenomenon. Every data tells a story and the story is of paramount importance for forming an honest opinion about a particular phenomenon. So what story is the UFO observational data telling? Our problem is to attempt to reveal this story.

1.3 Interest

This report is targeted to anyone interested in the UFO phenomenon.

2. Data

To answer our question we will use the UFO Sightings dataset [2]. This dataset contains the sighting report content itself including the time, location, duration, and other attributes in both the raw form as it is recorded on the National UFO Research Center (NUFORC) site as well as a refined, thanks to Tim Renner, standardized form that also contains latitude/longitude coordinates. We will use the refined form ( https://data.world/timothyrenner/ufo-sightings/workspace/file?filename=nuforc_reports.csv ) which has cleaned state and city names, standardized dates and latitude/longitude values of the cities in which the sightings occurred. Few things to note:

  1. All of the sighting reports are present - this is a complete scrape.

  2. The entire sighting report contents are present, including the full text of the sighting report.

  3. The sighting reports are geocoded at the city level. It isn’t a complete geocoding, but about 90k / 110k reports have an associated lat/lon.

  4. The “cleaned” version of the dataset has a few key standardizations applied like date times, city names, state codes, etc that make it easier to work with.

Columns information:

  1. summary: Summary of the report. Usually the first few sentences.

  2. city: The city of the sighting.

  3. state: The 2 character state code of the sighting.

  4. date_time: The date and time of the sighting in ISO 8601 (local time).

  5. shape: The shape of the sighting.

  6. duration: The duration of the sighting in no particular format.

  7. stats: Summary stats about the sighting (when it occurred, when it was posted, etc.).

  8. report_link: A link to the original report on the NUFORC site.

  9. text: The text of the sighting report.

  10. posted: When the sighting was posted to the NUFORC site.

  11. city_latitude: The latitude of the city in which the sighting occurred.

  12. city_longitude: The longitude of the city in which the sighting occurred.

[2].https://data.world/timothyrenner/ufo-sightings

3. Methodology

We will focus on the summary and text columns of our dataset since these columns contain the report information about the sighting event. Based on the information in these columns a list of words used to describe the UFO phenomenon will be constructed.

Specifically, some steps of Natural language processing (NLP) [3] analysis will be applied to this information. NLP is an area of computer science and artificial intelligence concerned with the interactions between computers and human (natural) languages, in particular how to program computers to process and analyze large amounts of natural language data. It is the branch of machine learning which is about analyzing any text and handling predictive analysis. Text cleaning or preprocessing in our case will include removing punctuations and other non-alphanumeric characters, numbers, stop words and words with the length <= 2 since all these entities do not help much in processing the information about sighting events and converting every remaining word to its lower case. After this preprocessing step we will calculate the frequency of the remaining words and create a list sorted in descending order by the frequency so that the most frequent words are at the beginning of the list (ufo_words_list in the notebook).

Using WordNet [4] we will identify nouns, verbs, adjectives and adverbs in this list. WordNet is a large lexical database of English. Nouns, verbs, adjectives and adverbs are grouped into sets of cognitive synonyms (synsets), each expressing a distinct concept. Synsets are interlinked by means of conceptual-semantic and lexical relations. WordNet is a useful tool for computational linguistics and Natural language processing. To get a visual sense of the most frequent nouns, verbs, adjectives and adverbs used to describe the UFO phenomenon we will create a bar plot of the 100 most frequent entries for each of these 4 word classes.

Next, stemming will be done which is taking roots of every word in the list. A bar plot of the 100 most frequent stem words will be created to get a feeling for the most frequent root words used to describe the UFO phenomenon. We will focus our attention on the top 50 stem words and create a word cloud to visually summarize the insights. We will sample sentences from the sighting reports containing one or more of these top 50 stem words to get an understanding how these stem words are used in the description of the UFO phenomenon.

Finally, using our list we will proceed to calculate the information entropy for each sighting report in the dataset. The information entropy quantifies how much information there is in a message. The intuition behind quantifying information [5] is the idea of measuring how much surprise there is in an event. Those events that are rare (low probability) are more surprising and therefore have more information than those events that are common (high probability). In other words, higher information entropy of a message does mean higher information content is in the message. The dataset will be sorted in descending order by the information entropy of the sighting reports and we will examine some reports with the largest entropy. These reports should be more surprising than the lowest entropy reports.

After all these steps we can attempt to reveal the story the UFO observational data is telling about the UFO phenomenon.

[3].https://www.geeksforgeeks.org/python-nlp-analysis-of-restaurant-reviews/

[4].https://wordnet.princeton.edu/

[5].https://machinelearningmastery.com/what-is-information-entropy/

4. Analysis

A Python 3 Jupiter notebook in the IBM Watson Studio is used to run analysis and produce results.

More information about notebooks in the IBM Watson Studio can be found here.

4.1 Basic Explanatory Analysis

Basic explanatory analysis shows that there is no difference whether we check for NaN or null values. The stats are the same. The report_link column is always populated. However, there are 54 rows where the text column value is not populated.

A quick glance at the summary column reveals that the summary column is always populated for these 54 rows except for the 54th row.

We see that there is one row in our data where both summary and text values are NaN. This row does not have any information about the sighting report so we will remove it. For the remaining rows we can use the summary value when respective text value is NaN. This can be accomplished by creating a new column called text_mining_info which will be the text when the text is populated, otherwise it will be the summary when the text is NaN. We will use the text_mining_info column in our analysis. Specifically, text cleaning will be applied to this column.

4.2 The NLP analysis - text cleaning

In order to run the NLP analysis we first need to install necessary libraries.

!conda install -c conda-forge  nltk --yes

Then we remove any number from the text_mining_info. We convert every word in the text_mining_info to its lower case. We replace all non-alphanumeric characters in the text_mining_info with a blank and keep only one blank between 2 words. Next, we split such modified text_mining_info on the blank and create a list of the remaining words. Essentially, we convert the information from every sighting report into a respective list of words. For each of these lists we remove stop words which are generally the most common words in a language, like words below.

explore_text_mining_info_df["mining_info"]=explore_text_mining_info_df["text_mining_info_no_non_alpha_array"].apply(lambda x: [item for item in x if item not in stop_words])

Finally, for each of the lists we remove words with the length <= 2.

explore_text_mining_info_df["mining_info2"]=explore_text_mining_info_df["mining_info"].apply(lambda x: [item for item in x if len(item) > 2])

At this point we have converted the information from every sighting report into a desirable list of words. A lot of lists were created, 112094 to be exact. Now we want to combine all these lists into one list. Unfortunately, a very convenient statement like this:

ufo_list =np.concatenate(explore_text_mining_info_df.mining_info2.values)

would cause the kernel of the notebook to die.

4.3 Dealing with dying kernel

It turned out that a simple divide and conquer strategy was enough to avoid the death of the kernel. So we split out respective dataframe into 10 smaller dataframes and create 10 corresponding lists and work with them.

We loop through 10 dataframes and for each dataframe we perform these steps:

  • concatenate respective words into a ufo_list variable

  • convert the ufo_list into a ufo_list_df dataframe, one word per row

  • create respective summary dataframe to keep words counts and their respective percentages of the total number of the ufo_list_dt rows.

  • keep only words with percentages greater than 0.001 percent.

0.001 percent cutoff is determined by trail and error. It is the first value at which the kernel started dying, meaning we could not use lower order magnitude values for the cutoff. These 10 summary dataframes are concatenated and respective words counts and percentages are recalculated since some words appeared in more than one dataframe (see the notebook for details). We also added for each word respective probability (prob) which is simply perc_tot/100. Below is a snapshot of the head records of the sorted final dataframe by the count in descending order.

4.4 The list of words used to describe the UFO phenomenon

The list of words used to describe the UFO phenomenon is the info column of the above dataframe. First 30 most frequent words in the list are displayed.

We see that the first few words suggest some bright moving object in the sky and light, lights as the most common words used to describe the UFO phenomenon.

4.5 The NLP analysis - Identify nouns, verbs, adverbs and adjectives in the list

As was mentioned in the Methodology section we use WordNet for this task and find that most of the words in the list are nouns. Interestingly, there are 931 words that are classified as ‘None’.

noun    4677
verb    1395
None     931
adj      561
adv      299

If we list these ‘None’ words we can easily see that ‘would’, ‘could’, ‘since’, ‘else’, …, etc. are in this group so ‘None’ words in essence do not carry information important for the description of the UFO phenomenon and we remove these words from the list.

4.6 The NLP analysis - Stemming

To take main stem of each word in the list a PorterStemmer object is created and used.

Stem words are summarized by counts and percentages.

The bar plot below of the 100 most frequent stem words shades a light on most often root words used to describe the UFO phenomenon. It seems all action is in the first half of the plot with ‘light’ as a very distinguished leading stem word.

4.7 The information entropy

Finally we calculate the information entropy for each sighting report. Each word of the sighting report is searched in the dictionary that contains words and respective probabilities

ufo_dict_in = ufo_summary_stats_word_type[['info','prob']]
ufo_words_dict = dict(ufo_dict_in.values.tolist())

and if found respective probability is added to a ‘score’.

explore_text_mining_info_df['score']=explore_text_mining_info_df['mining_info2'].apply(lambda x: [ufo_words_dict[item] for item in x if item in ufo_words_dict.keys()])

Essentially, each sighting report is converted into a list of probabilities and using this list respective entropy is calculated as

explore_text_mining_info_df['entropy']=explore_text_mining_info_df['score'].apply(lambda x: sum([-item*np.log2(item) for item in x if item > 0]))

which is

formula where the log function uses base 2. Sighting reports are sorted by the entropy in descending order. Out of curiosity, the summary information for the top 6 sighting reports is displayed below.

Certainly, the third and sixth are quite intriguing.

5. Results

Bar plots below give us a visual sense of the most frequent nouns, verbs, adjectives and adverbs used to describe the UFO phenomenon for the 100 most frequent entries for each of these 4 word classes.

nouns

Clearly, ‘light’, ‘object’, ‘lights’, and ‘sky’ are leading nouns. It is interesting ‘saw’ and ‘see’ are classified by WordNet as nouns although we would expect for ‘see’ to be a verb and for ‘saw’ to be past tense of see.

verbs

The leading verbs are ‘moving’, ‘seen’, and ‘looked’.

adverbs

‘around’ is the top adverb while ‘bright’ is the top adjective.

adjectives

To make it easy to spot frequencies of the top 50 stem words used to describe the UFO phenomenon we created a Word Cloud.

We also list the top 50 stem words.

light object sky move look saw like see bright time shape 

appear seen wit star white sight craft second night red 

back two minut note around fli watch first seem went 

notic orang disappear ufo north plane sound south east report 

west thought away direct color flash approxim travel speed

In order to gain understanding how these root words are used to describe the UFO phenomenon we did sample sentences from the sighting reports containing one or more of these words. This wasn’t anything fancy, we just limited our pool to the first 10000 reports and focused on the reports with the largest numbers of sentences in which these words appeared.

Here are the links for some of the reports:

http://www.nuforc.org/webreports/010/S10250.html (bright surrounding light,banging on car, animal life affected) http://www.nuforc.org/webreports/011/S11956.html (abduction to underground facility)

http://www.nuforc.org/webreports/011/S11676.html (two bright objects perpendicular to one another)

http://www.nuforc.org/webreports/014/S14868.html (one bright object which turned into eleven)

http://www.nuforc.org/webreports/011/S11868.html (teardrop shape)

http://www.nuforc.org/webreports/133/S133485.html (changing shape)

http://www.nuforc.org/webreports/132/S132745.html (fluid formation flying disks)

http://www.nuforc.org/webreports/133/S133288.html (observed green star)

http://www.nuforc.org/webreports/009/S09604.html (craft floated and made no noise)

http://www.nuforc.org/webreports/004/S04806.html (5-8 smaller objects began to fall out of the larger object)

http://www.nuforc.org/webreports/004/S04195.html (if you look at it’s lights it can affect you emotionally)

http://www.nuforc.org/webreports/006/S06339.html (light in the sky that changes color)

http://www.nuforc.org/webreports/003/S03372.html (memory gaps)

http://www.nuforc.org/webreports/002/S02047.html (drain the power from the battery of car)

http://www.nuforc.org/webreports/003/S03291.html (S-pattern)

http://www.nuforc.org/webreports/011/S11772.html (not from this world!)

http://www.nuforc.org/webreports/010/S10553.html (thousands report UFO in Papua New Guinea)

http://www.nuforc.org/webreports/132/S132430.html (glowing light that moved, stopped, glew brighter etc.)

http://www.nuforc.org/webreports/007/S07673.html (loop over each other)

http://www.nuforc.org/webreports/008/S08795.html (the light just sucked up in nothing and vanished)

http://www.nuforc.org/webreports/004/S04967.html (it hovered for a few seconds and then disappeared)

http://www.nuforc.org/webreports/003/S03298.html (“no plane I know can do this”)

http://www.nuforc.org/webreports/008/S08101.html (the CD mysteriously played backwards!)

http://www.nuforc.org/webreports/005/S05196.html (zig zag)

http://www.nuforc.org/webreports/005/S05813.html (THERE WAS NO SOUND COMING FROM THIS OBJECT)

http://www.nuforc.org/webreports/008/S08439.html (flashing lights)

http://www.nuforc.org/webreports/009/S09627.html (intense UFO encounter sites)

http://www.nuforc.org/webreports/011/S11867.html (object travels at great speed, suddenly changes direction)

http://www.nuforc.org/webreports/013/S13847.html (extraordinary maneuverability and speed)

 

The reports suggest following fascinating fragments of the story the UFO observational data is telling about the UFO phenomenon:

It is about a moving, bright object or objects in the sky with distinct light features, extraordinary, unusual, flying characteristics, power to affect nearby things or living beings and some noticeable, sometimes, changing shape.

The object can look like a colorful light in the sky that changes color, for example, from red to green to yellow, the light can be flashing or the light that disappears or the object looks like a moving star. For example, “glowing light that moved, stopped, glowed brighter and dimmed out” or “right before it would have hit plane, the light just sucked up in nothing and vanished”.

It seems if you look at the object’s lights it can affect you emotionally!

These objects can hover and disappear, move erratically, like zig zag, can float and make absolutely no noise or sound.

The object can be a silent airborne object and can come to a complete, sudden stop and smaller objects can fall out of the object.

These objects can fly in a perfect formation like a perfect line or in a strange formation, for example S-shaped pattern, can loop over each other and display extraordinary maneuverability and speed and can suddenly change direction.

It looks the object can somehow affect the memory of a person looking at it, causing memory gaps!

Besides causing electrical failures in the car, wipers beating windshields, etc. the object can affect animal life as well.

It looks when these objects are around, CD players and radios can mysteriously play the CD backwards!

The objects can drain the power from the battery of car, seems to hear what a person was thinking and entities can abduct the person.

The most frequent 5 shapes for these objects are light, circle, triangle, fireball or unknown.

Sometimes many people witness these objects (see Thousands report UFO in Papua New Guinea) and some areas are intense UFO encounter sites.

These things look not from this world!

Finally, the links below are for the top 6 sighting reports based on the information entropy that we listed in 4.7 section:

  1. http://www.nuforc.org/webreports/002/S02047.html

  2. http://www.nuforc.org/webreports/086/S86119.html

  3. http://www.nuforc.org/webreports/117/S117952.html

  4. http://www.nuforc.org/webreports/072/S72574.html

  5. http://www.nuforc.org/webreports/116/S116501.html

  6. http://www.nuforc.org/webreports/109/S109660.html

The third report is about a Close Encounter of the Fifth Kind which Wikipedia defines as “a UFO event that involves direct communication between aliens and humans” [6].

[6].https://en.wikipedia.org/wiki/Close_encounter#Close_Encounters_of_the_Fifth_Kind

Some captivating parts of that report:

The disk was still flying toward us, slowly, maybe 20-mph. It flew the same way over the canal until about 300-yards away. Then it went gradually higher to about 400-yards, almost over head. It’s changing direction slightly now, moving to the opposite edge of the canal and over a point about 100-yards in front of us. It slowed, descending as it went by. It’s absurd not being aware that it’s a UFO but it never entered my mind. A UFO is exactly what we saw, the real thing. I never saw one before and this one is really close. It didn’t get closer than 200-yards to where we stood. As the craft flew by, it was perfectly circular, about 35-feet in diameter give or take five. At that height it looks bigger than, say, a full moon rising on the horizon and it’s brighter. As it moved it made a bright white glow with a hint of yellow, bright but without a strain… easy on the eyes….

Here’s what happened. We watched as the craft came lower, stopped, and became darker. Two small lights appeared. The craft was parked in the air just inside the levee of Big Lake maybe as close as 200-yards away. As we watched, something hit me, a marvelously happy feeling and I knew at the time it was coming from the craft. I thought THIS is a really good sign whoever they are….

The happiest feeling of anticipation had suddenly come upon me. I knew to expect something wonderful to happen between them and me. They were doing something to affect my mental state giving me a wonderful feeling of euphoria. It‘s nothing like anything I’ve ever felt, a totally joyful happy feeling beyond description. I was so glad they were here and understood, waiting for something. I dared not move a muscle although knew I wasn’t restrained. Whatever came next I wanted to be part of it. Next, I was sensing a projection of thought to me, when soon thought-speech appeared in mind in my own idiom….

The most bizarre part of the 6th report is a description of a possible extraterrestrial biological entity:

Once back on the interstate I drove for quite some time when finally I just had to pull off. By this time it was after midnight. I took a well lit exit and saw a Wendy’s but it was only open for drive through. All of the gas stations were closed. All the while I’m checking the skies. So, I got back on I-77 and looked for another exit. That’s when I saw a Shell station sign and took that exit. It was on the left side so we had to go over the overpass. We could tell it was closed as soon as we came over the overpass. All of the parking lot lights were on but it was closed. I was in pain because my bladder was hurting now. I pulled partially into the entrance of the Shell station and then made a u turn. As I completed the u turn we saw a guard rail right in front of us.

At the very end of the guard rail as we were swinging around we saw something standing there. I couldn’t take my eyes off of it. It was right next to the road. Right next to the door of the car. It seemed to be! bristling. Like it was luminescent in some way, but it was really darkly colored like it was camouflaged. I actually thought it was someone dressed up in that camouflage netting stuff when it moved towards the glass right next to my daughter. She put her hands and arms up over her face and head and tucked her head downwards. That is how close that thing was to her door. I honestly thought it may have been a poacher until I could make out the shape of it. I could see a larger than normal rounded top of a head then it came down to a short little neck area then straight out to both sides to shoulders. (Human shoulders aren’t shaped that way.) Then it’s arms dropped straight down. The thing that made me realize that it wasn’t human was mostly that it’s waist was only like 8 inches wide and that it was about maybe 4 feet tall. I could see from the top of it’s head to its hip area through her window. I remember thinking “It’s so skinny!” Next thing I know we are past it and my daughter is asking me, “Was that a person?” I remember looking back over my right shoulder through the window behind her and still seeing the thing. It was moving and I had never been so terrified in my entire life.

My daughter said that she didn’t look at it but then moments later she said that it was really thin. I remember asking her how do you know that if you didn’t look at it? She still can’t answer that question, but is adamant about not looking at it.

Besides a possible abduction in both cases, what is so striking about the 6th report is that points to a possible occurrence of missing time:

Strange thing is that the clock on her cell and the register receipt didn’t match. The receipt said we left there at 10:05 and my daughter had gotten a call from her husband while we were at the table and the time read 10:46. That would mean that by her cell we left there after 11:00 PM. I know we weren’t sitting there for that long. I called the Denny’s last week and spoke to a manager to see if the register’s time was off. But, it was accurate. I can’t explain the difference.

Feel free to click on the links for these 2 reports or any of the above reports to explore them.

6. Discussion

In his book [7] Allen Hynek established beyond reasonable doubt that UFO cases are not all misperceptions or hoaxes. Publicly available datasets about the UFO observations allow us to explore freely the sighting reports not for the sake of their story value, but for the insights what people have actually experienced that witnessed these things. Our problem was to reveal a story the UFO observational data is telling about the UFO phenomenon. We combined two ingredients, a free available UFO dataset and data science to unlock insights about the UFO phenomenon. In other words, we did focus on the narrative derived from the data and analysis based on the Natural language processing, stemming and calculation of the information entropy. Essentially we did build a list of words used to describe the UFO phenomenon along with their respective frequencies and we did focus on the 50 most frequent stem words. Using the information entropy to quantify how much information there is in each sighting report and looking at the top sighting reports in the information entropy we did reveal some aspects of the UFO phenomenon, for example abduction and occurrence of missing time, that were not captured by the 50 most frequent stem words. For example, stem words related to abduction and occurrence of missing time are abduct and clock and they are on 1239th and 559th place respectively in the list of the stem words.

Due to the dying kernel, with our simple divide and conquer strategy to avoid the kernel death, we were only able to keep words with percentages greater than 0.001 percent in the list (see the section 4.3 for details). This means after all 10 splits were processed we were able to include only about 17% of all words in the list as it can be seen from the stats below

This means that we were able to uncover only fragments of the story, but even these fragments will definitely capture the imagination of readers. To capture the full story the UFO observational data is telling about the UFO phenomenon additional methodology and analysis is needed to process all words, for example MapReduce approach in which the work is divided into smaller parts. Each part is processed independently in parallel on the cluster of servers to get individual outputs. These outputs are further processed to get the final output. Building a complete list of words used to describe the UFO phenomenon along with their respective frequencies is the word count MapReduce problem that can be easily solved in the MapReduce framework (word count). Our gut feeling is that we should not expect significant changes in what we uncovered so far by including low frequency words (0.001 percent or less) in the list.

Even with the fragmentary information extracted by our approach we can see that the full story is going to be incredible, about things that look not from our everyday world. The story will be about moving, bright object or objects in the sky with distinct light features like a colorful light in the sky that changes color or like the light that just sucked up in nothing and vanished.

Their flying characteristics are unusual, extraordinary. These objects can hover and disappear, move erratically, like zig zag, can float and make absolutely no noise or sound. For example, the object can come to a complete, sudden stop and smaller objects can fall out of the object. Or the objects can fly in a perfect formation like a perfect line or in a strange formation, like S-shaped pattern, can loop over each other and display extraordinary maneuverability and speed and can suddenly change direction.

The objects have power to affect nearby things or living beings. For example, it seems if you look at the object’s lights it can affect you emotionally! Or it looks the object can somehow affect the memory of a person looking at it, causing memory gaps! They can affect animal life as well and drain the power from the battery of car, cause wipers beating windshields and it seems they can hear what a person was thinking. It looks when these objects are around, CD players and radios can mysteriously play the CD backwards and they can cause occurrence of missing time! Finally, entities in these objects can abduct the person.

[7].Hynek, Allen J. (1998) [First published 1972]. The UFO Experience: A Scientific Inquiry. Da Capo Press. ISBN 978-1-56924-782-2.

7. Conclusion

The purpose of this project was to attempt to reveal the story the UFO observational data is telling about the UFO phenomenon. This story would capture the insights what people have actually experienced that witnessed these things. We did show that focusing on the narrative derived from the UFO observational data and analysis based on the Natural language processing, stemming and calculation of the information entropy we were able to uncover fascinating fragments of the story. Information extracted by our approach so far points to the story about things that look not from our everyday world. Future directions include researching how to incorporate additional methodology and analysis, for example MapReduce, to capture the full story the UFO observational data is telling about the UFO phenomenon.

Related Posts