Abcruz:



You're totally correct about the skepticism regarding the adoption of AI in medicine. Some doctors are strongly against it because of the threat of losing their professional relevance through automation. On the other hand, people are scared of entrusting their health challenges in the hands of a machine who has no empathy or emotion and probably designed and deployed by someone who isn't medically grounded to recognize misdiagnosis of the system.

Nonetheless, medicine is a very complicated field of endeavor that cannot be practiced without the human touch. It will be in the best interest of both doctors and patients if AI is designed to automate routine clinical practices while the doctors focus more on care based medicine.

Well said. Handling of routine clinical practices by intelligent machines will greatly reduce heavy reliance of many patients on a single doctor (talk of patient doctor ratio) most especially in countries like Nigeria.

Enceladus:




Absolutely fantastic... [b]The CNN (?) [/b]might even find clues and patterns oblivious to us that might help it identify these issues perfectly.



I focus on NLP mainly also, and I think they might be the last nut to crack before we get artificial general intelligence. Like for example GPT-3 is showing some signs it can reason and decode some quiet complex natural language and it understands a statement in multiple languages. I once started a project like that, but the sheer volume of the effort required made me abandon it sharply




Absolutely. No hands-on coding while learning == Binge-watching Netflix.

Yeah, CNN. Come to think of it, despite the tremendous success deep learning has achieved in medicine, clinical experts are of the opinion that they can not rely on a technology whose decision making process is not evidence based (and this is what standard medical practice is all about).

Given a vast amount of training examples, neural nets are very good at extracting features and making precise predictions that can even surpass human performances in certain cases. These remarkable achievements are quite fascinating but there's a problem; we do not know how these features are being extracted given the black box nature of neural networks. This is a major concern for health care experts and it is quite debilitating to the adoptability of AI in medicine.

As for NLP, frankly speaking, I enjoy doing it more than CV and I hope to take some advanced courses in it to gain more knowledge about the field.

GTP-3 is a monster but we are yet to see a truly intelligent system. Anyone to come up with that will control the world for the next 50 years because so many industries will be disrupted and countless number of businesses will go into extinction.

What NLP library/libraries do you enjoy using most? Spacy and Textblob top the list of my own favorite libraries. Also do you process any other language apart from English?

Semtu:


Alright, I would love to see the curriculum also, let me see how well it was broken down.

You can find it here

Semtu:


Wow, nice to see schools are finally delving into tech. Well, rather than going into AI for primary kids. You have to first start by teaching them python programming, make python fun for them by writing simple game scripts and such. I don't think primary school kids are ready for AI. As for the Robotics aspect, you can play around building simple robotic toys.

Thanks for your input. Just this evening, I came across a curriculum developed by MIT just for kids. I think I'm just going to use that.

Enceladus:
Conclusion

The experiment was a bit of success although the agent didn’t maximize its potential mainly because it can only sense the board only through the region around it’s head and a better option is the snake seeing the whole board as an image through a convolutional neural network. The agent was able to eat about 50 apples at max.

For the next experiment, I will be training an agent to play the old shooting game CONTRA and I am really excited about the training. I will also build upon the lesson learnt by this experiment by giving the agent more awareness of the environment by providing the agent the direct image render of the screen. IF the next experiment is successful, other prospective games for experimenting include Asphalt, Mortal Kombat etc. and also maybe soccer (but definitely the old versions like winning eleven 2008 and will probably need the help of a C++ dev to tune the game to a usable environment)
Adios guys.

This is amazing!

I also use Pytorch for deep learning. I recently trained a neural net model and wrote a utility function that can help eye care specialists (ophthalmologists and optometrist) differentially diagnose certain eye conditions (conjunctivitis and uveitis) in real time. I achieved an accuracy of close to 90% but I still need something higher than that. My ultimate goal is to build a stand alone system where the model can be embedded and used by clinical experts.

I also find NLP very interesting and I have developed, debugged, tested and deployed a keyword generating API that extracts important terms from a text. Currently, I'm working on a Chabot using Rasa. I want it to be able to hold reasonably conversations regarding eye health and make tentative diagnoses based on users' intents.

Data science as a whole is very interesting but I don't think anyone can go far by just watching YouTube videos and taking some courses online. Personally, I combine PDFs with videos and do a lot of hands-on.

Semtu:


AI is not just a single field one can abruptly talk about.
Well, the branch of AI I've dealt with are Machine learning and Robotics. These are two interesting fields. Machine learning on the one hand basically deals with machines or systems capable of learning from test data. After thorough training, the machine should be able to carry out accurate predictions be it categorical, regression, and lots of other methods. A major disadvantage of machine learning is the high susceptibility to error. For example, its very easy to over fit a Machine Learning model if care is not taken.
Robotics on the other hand is the use of mechanical and electronics systems to develop machines capable to carrying out tasks with little or no human intervention. Robotics has numerous disadvantages ranging from high cost of implementation to the need for constant maintainenace.
Other fields of AI includes deep learning and NLP. Deep learning has to do with learning across multiple layers of training. They generally produce systems with higher accuracy than machine learning . In neural networks, weights are adjusted over several layers till your error rate is small enough. It's somewhat similar to deep learning.

As an AI enthusiast, I suggest you read up about the various fields of AI before delving into it. In the real world, there is no specific field like AI. You have to focus on one of the numerous branches of AI. That said, your language of choice should be Python. Python is the language of AI.

What a good analysis!

Do you have any idea where I can get AI and robotics curriculum for primary school kids? I'm currently in talks with a school proprietor who wants me to introduce his kids to hands-on AI and robotics.

Doradorwa:
I av stopped using my glasses since I got back from school due to the nature of my job... I recently discovered that my eyes are getting to watery hence blurring my vision... The eyeballs are so heavy hence making me have headache. I can't walk in the afternoon in peace. But like I said since 6 months or so I av not been wearing my glasses.... There is the fleshy part in the eyes...it get so itchy... I don't know... I will still need to see a doctor but work has been hectic

You should see a doctor ASAP.

Doradorwa:
I was not with her during her eyes examination... but I might get my eyes checked too.... I recently got a correctional eye glasses for myopia but I've been having difficulty lately. I believe there might have been some mix up with my own eyes examination hence giving me worries that the cataract might be genetic. But thanks anyway.... I'm grateful

With or without family history of cataracts, people are still likely to develop the condition as a result of aging. Although family history is an important risk factor. There's no need to panic, just make sure you go for your checkup regularly.

What issue are you having with your glasses if I may ask?

Doradorwa:
Seriously I have no idea... is it strange if cataract is in the both eyes? Tho we recently got one of the eyes operated on... just one more to go...

Bilateral cataracts is not uncommon. With a comprehensive eye examination, her doctor should be able to determine what the etiology is and expectedly relate it with you. This is very necessary in situations where the condition is genetic.
Between, I'm glad to know that one eye has been operated on and I hope she's experiencing a good improvement in her vision (visual acuity).

Doradorwa:
Thank you sir

You are welcome. Just wondering what the etiology could be for her to be having bilateral cataracts.

Doradorwa:
Good evening everyone! Pls im begging everyone on this forum that knows any foundation around ogun or lagos state that help people with cataract for their operation free of charge to help me help my mom. Pls help me. We dont have the money the hospital is demanding. I told her to use my school fees but she said no. She said if she use my school fees she has no source of getting it back. She is the only one i have. The thing has consume her eye. Pls help me if you know anyone that can help me. Pls help me im begging anyone and everyone who can. Pls call me on 09058380414 for any recommedation.

Try to locate any Tulsi Chanrai foundation close to your area. Their center in Kebbi state offers a very affordable cataract surgery (less than 20k) with a high level of precision.

Harkstetunz:
Hello guys! I've been an ardent visitor of this section for a while now and believe me guys, you're all doing well .
I literally haven't been fortunate to come across any thread detailing Artificial intelligence(there might actually be many) ; maybe you guys do not have penchant for it.
Please, bosses, help about the things you know about AI (pros and cons included) and the things AI enthusiasts should know and be prepared for.
Thanks.

Artificial Intelligence is a vast field, you should be specific about your questions.

Available247:
Anytime I abstain from any sexual activity for a period of time i always encounter wet dream i.e ejaculating when sleeping.

Sometimes it can be a sex dream, stressful

dream or wouldn't remember anything at all. They say wet dream is part of puberty, but I'm no longer a child, I'm a full grown adult and wet dream irritates me a lot. Cleaning up my clothes and sheets because of it can be a pain in the neck.



Does anyone here still encounter wet dream after abstaining from sexual act for a long time and does it have any effect on our mental wellbeing

You just described what I'm currently going through. I'm 26, no pornography, no masturbation and I still go through this frustrating situation, at least two times a week.

LordFriar:
If you've ever had this trouble, I need your help.

Bg Story: Anaconda keeps returning 'Failed to create Menus'' whenever I've it reinstalled.

I've tried out the suggestions from Anaconda itself .. nth +ve. Help smone

Go to Anaconda download archive and grab the 2019.03 version. I encountered a similar issue while helping someone to install it and the solution I just provided worked for us.

haiphy:
pls can anyone help with research result analysis? holla me @ 08039522122

Send a me a mail at ibromodzi@gmail.com

BelieverDE:

I like you, brother!

I suppose I'll have to focus more on the Shapiro Wilk test. There's no need exploring the K-S test when it's sensitive to extreme values. I could many rows of data and I wouldn't want to be scared that extreme anomalous values can affect my result. Thanks a lot, brother.

Cheers!

You welcome bro.

BelieverDE:

My God!!!!
When will I finish learning everything?!

I read this post in the morning and decided to make more research before asking for more clarity, but I became confused the more.

Different people on researchgate were proffering different answers on the best normality test. Some said when we have a sample size that is less than 50, we should use the Shapiro Wilk's test; and if the sample size is greater than 50 we should use the Kolmogorov-Smirnov test. Then different researchers began talking about sensitivity of the two tests to confuse me the more.

To me, if the population is greater than 50, why perform another test when you know the population would be normal (thanks to the Central Limit Theorem) ?

About two researchers went basic and talked about the use of histogram. I've got no time for that, and would rather use the test for skewness and kurtosis instead.

You must come from a Statistics/Mathematics background, what course would you suggest for me if I want to perform a two paired T test?

If the sample size is above 100, would you still use the Shapiro Wilk test? If no, why?

That's the power of research. The more you read, the more information you have at your disposal.

Visual inspection of data distribution (histogram, box plot, q-q plot, p-p plot) is informal, unreliable and does not guarantee that the distribution is normal. Nevertheless, presenting your data visually gives your audience the ability to judge the distribution themselves.

There are many formal tests that can be used for normality aside Shapiro Wilk and K-S tests. However, the former is recommended by many researchers regardless of the sample size, owing to the fact that K-S test is sensitive to extreme values.

You can read more about Normality test here https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3693611/

@ the emboldened.
I'm a final year student of optometry and vision science. No statistics/math. Although, I took some basic math/stat/CS courses in 100/200L

wisemania:


Alright

Share if I'm not the only one lost here.

Let me know what you guys know

I've been found wanting in statistics for a while, please how can you help me?

Is there a practical udemy course that explains these concepts in-depth?

Lol.... You can try www.pythonfordatascience.org

BelieverDE:

Thanks, I'm glad I am on the right track.

Let's assume the population of a T test, after checking for kurtosis and skewness, is not normal. What alternative test do you think can be carried out for hypothesis testing?

While skewness and kurtosis are supported for testing normality by some texts, the most acceptable standard is by using a formal test (i.e Shapiro Wilk's test) or visually by using q-q/p-p plots.
That being said, an alternative, non parametric test for a paired T test will be Wilcoxon paired signed rank test .

BelieverDE:


You have spoken well. Thanks for your verbose explanation, I really picked some insights from it.

I would like to ask; at the emboldened part:

a. In the example, I suppose a two paired hypothesis test would give us the answer, right?

b. I couldn't think of anyway of statically analyzing the data. I guess that's where Machine Learning come into play since we want to predict an outcome


Are my answers correct?

I'm glad you were able to pick some things from it. Frankly speaking, I asked myself if I wasn't typing rubbish while trying to come up with a simple, yet wordy answer to the question.

Pertaining to your question, yes your answers are correct to some extent. I like the way you think.

In the first example, the clinical trial will most likely involve testing the new drug against a placebo on the same set of patients, in which case a paired T test (or alternative non parametric tests if assumptions of normality and homogeneity of variance are not satisfied).

The second example is a predictive analytics and that's where ML comes in - Logistic Regression is a good algorithm to start with.

KlausMichaelson:
Good day House.

I really need someone to enlighten me in the real application of Data Analysis in real life. I mean, I actually know how to provide a good Visualization for any Data, I can use the different functions in Excel to an extent. I can use the Statistical Analysis part of it especially when trying to analyze a typical student's project.

But my Question is, "is that all?"
Come to think of it, if I actually get employed as a Data analyst, Do I just need to provide a Visualization for the company to see how far different factors(Demand, supply bla bla) are playing out at the end of the month?

I honestly don't get the whole thing. The only application I have come to Realize so far is based on Visualization and Statistical Analysis only. Or is the Statistical Analysis all that matters? Things like ANOVA, REGRESSION, CORELATION etc??

Please can someone tell me the real application of Data Analysis in real life. I'm very much confused.
Thank you.

From your write up, I can deduce that two things are missing in your data analytics armament. First is the theoretical knowledge of data analysis (most guys joke with this always) and the second one is the domain knowledge needed to navigate through any data you are working on. You can't make any meaningful conclusion about datasets without these two.

Now, back to the question. Honestly, a simple google search would have given you the answers to your questions, but I'll still try to break it down according to my own little understanding of data science.

Any data collected in its raw form has no meaning until it is processed according to what the company requires so as to utilize the data solely for decision making purpose (making conclusion), which in turn helps the company to grow. Now, there are two important things here,
1. The company's requirement (i.e why was the data collected in the first place?) - Let me give you a couple of real life examples here:
a. The government wants to know whether a newly developed drug will be effective in the treatment of Covid-19. So they do this by conducting a round of clinical trials on patients willing to participate. Remember why the clinical trials were conducted? To test the effectiveness of the drug.
b. A financial institution (bank for instance) wants to develop an advanced system for detecting fraudulent transactions among their customers. They do this by getting a large record of transactions that have taken place in the past. Again, you remember why they needed the record of past transactions? I hope you are getting the gist by now?

2. Now, let's continue with the second important thing in data analysis process which is the goal/objective of obtaining the data in the first place - to make decisions or conclusion about the data.

Going back to the examples I gave above, the government can conclude whether or not the drug is effective based on the outcome of the clinical trial, the financial institution can spot fraudulent transactions based on the model built with the data collected, a country can project their future population based on the available data and make appropriate policies to address any resulting economic/social challenges, Netflix can recommend movies for you based on your history, this makes you stay glued to them, which eventually generates more revenue for them, YouTube suggests videos for you based on your past activities, you are tempted to watch more which means more money for them.

So what's data analysis in real life all about? Three basic things;
1. Why collecting the data?
2. What conclusion (actionable insights) are we making from the analysis?
3. How does the conclusion affect the growth or otherwise of the company?

The takeaways...

When carrying out any analysis, ask yourself or the client these questions;
1. What is the objective of the analysis?
2. Do we have research questions we are trying to answer?
3. Do we have any research hypothesis (which eventually leads to conclusion) ?

The above are what your visualizations and statistical analysis tell you. The more experienced you are, the more accurate you are likely to get the steps.

I hope you find this useful and grow with it.

cochtrane:
We should throw some challenges here from time to time. Get everyone spiced up on some topics in data science, R, machine learning concepts, bayesian statistics, etc.

Nice suggestion. Been a while boss.

Datst136:


Please is there a clear distinction between thesis analysis(inference statistics) and data analysis??
I believe there is a form of correlation between the two.
Techniques like ( correlational , regression, ANOVA) I believe are synonymous to the two.

There's really no clear distinction between the two. Thesis analysis involves the use of both descriptive and inferential statistics, and in some cases, predictive analysis (which are types of data analysis). Therefore, thesis analysis is realized by using different data analysis types/forms.

TheGenius001:
I hardly see anyone talk about SPSS on this thread. Why is that so??

SPSS is majorly used in academia and I don't think most folks here are into project/dissertation/thesis analysis which mostly deals with inferential statistics.

Personally, I have not seen a reason to use SPSS. Excel, with it's various add-ins can give me what I want.

Twirelex:

I did an analysis on the nairaland data you scrapped... it's on my blog
https://talkrtive.com/post/text-mining-with-r/

thanks for sharing the dataset

Nice work bro. I believe a higher accuracy can be achieved by implementing more sophisticated information extraction techniques such as tfidf, word2vec, doc2vec and RNN. You can as well work around the imbalanced dataset by using SMOTE.

talkcash:
[b][/b][i][/i]Programming can be easy, hard, frustrating and confusing but if you learn these languages as listed you will be on track.


1). HTML
Html is the simplest language to learn. It is mostly used for defining other languages that run on web.

2). CSS
This comes after HTML, it is used to beautify a html code, that is why it is called cascading styling sheet.

3). JAVASCRIPT
Javascript is a language that is used for client side scripting (OOP scripting). This language is among the top three best languages among the history of programming, it can do almost everything in the back end area, plus it is easy too.

4). Python
Python is a server side language, it is good for pretty much everything but is a little bit difficult to learn. Python is good in both web development and android app development

5). Ruby
Ruby is a language used for web development ONLY, it is a pretty old language but some people use it.

6). PHP
My language, PHP is also a server side language just like Python. It can partially be used for mobile app development.

7). JAVA
Java is used for developing android apps, and will soon fade away, many apps are developed with java.

. Kotlin
Kotlin will be the language to take over JAVA, because it is also used for app development, and is user friendly too.

9). Swift
Swift is used for making IOS apps that is apple apps, to learn swift you have to have experience of java and Kotlin.

10). C++
This is used for developing windows applications, it is pretty difficult, TRUST ME!

You need to make some basic findings before putting up stuffs on a public forum like this, to avoid misleading newbies. There are many errors in your post but I'll quickly point out the obvious ones.

First and foremost, HTML and CSS aren't programming languages. Get the difference between a markup language and a programming language.

JAVASCRIPT is not just for client side, it is used for both front end and backend.

Python is not just a server side programming language, in fact, backend is just one of the many wonders Python can do. It is a general purpose language.

Lastly, C++ is not just used for desktop applications. It is used in writing OS, DB etc..

ekrizz:


Evening boss. Please, how can ML/DL models be turned to web/mobile apps? Do you have any ebooks or video on it?

Evening! For web applications, you have to learn a Python framework (Flask is relatively easy compared to Django). After that, read up Pickling. If you don't know how to start, just google how to deploy ML models using Flask and Heroku. There are many videos on YouTube.

Note: You don't have to master the key concepts of any framework to have your models running on the web. With less than 100 lines of code, you can deploy your models.

You may as well use JavaScript if you are familiar with it.

Oddy16:
You may still not find it interesting because they were a bit fast in the video. I think they were trying to cover up as much topics as possible. What I normally do is to write down the topics and summary of what it is all about, then search YouTube for where they digested any topic I want to learn well enough.

Here is the link,

https://www.youtube.com/watch?v=RdTozKPY_OQ

Thanks

Jas80:


It works fine, but consider a scenario where the user left on day 2, that's on Tuesday and spent just 1day, this scenario will lead to a situation of 3%7

The result does not change. 3%7 will return 3...

Oddy16:
It's funny how you listed all of these for people to learn while you have problem with the little I listed. And you do not get it that I am a bookworm cos you don't know whether I am or not.

Someone asked about ML and I explained to him how he can start up with that. You don't even know whether he/she can do analysis already, or interested in the "sharp sharp" you are advising.

Actually, the intro to statistics can help him greatly in analysis. You should never advise someone to dim his/her curiosity no matter what. If he/she is curious about anything, direct him to the right path. He/she has already seen what is needed for data analysis and not here saying ML doesn't make sense in Nigeria.

The interview you attended, if someone else interviewed for that position and has the same skills as yours and maybe has a knowledge of regression analysis, who do you think will be picked?

The ML you are talking about that doesn't make sense, is it not with the same knowledge that some Nigerians are winning some money on Zindi Africa? Zindi Africa is supposed to be for Africans but guess the country that is winning mostly there?....

Not everybody here is sitting at home hustling for data science and analytics jobs. Some have good jobs and want to transition, and it will be really good if they have a very good understanding and be really good at it.

I became very good with Python because I always download and go through notebooks on Kaggle. Why did I go to Kaggle on the first place? Because I was curious about ML. Nobody can do ML without processing the data first and you get to see notebooks that are just for analysis.

I couldn't learn Tableau even after watching videos till I went back to statistics to really understand about data. I remembered asking here which videos I can watch to really understand it. Do you also think that analysis is just about dragging and dropping and drawing charts? Won't you recommend or do some predictions for the company based on what you understood from the data given to you.

Oh, before I submit, let me also said that curiosity about ML also led me to learn about softwares too. I can build websites also. While I have come up with a portfolio on data science and applying for the jobs, I am also building a web project to apply for front-end web developer.
Yes, I was curious about ML, and it has led me to acquire more than what I expected.

ML Engineers are also available remotely. You should be curious enough to learn what is relevant beyond your location. Upwork and Fiverr are all global. Someone has brought two questions here from upwork and we were just confounded. A work on excel actually, and we couldn't get it. All of them are available online. The day I watched Edureka 4hrs video on excel, I was dumbfounded to know so many things I can do with excel. You can build bots using excel, finance goal seeks, VBAs, Macros...

We are still here talking charts and functions and still asking what excel can do that pandas can't do. If someone says anything that is foreign here, we try to shut down the person with "it's not applicable in Nigeria bla bla bla". Do you even know the worst thing about us youth in Naija? We are even ready to explore or think beyond our locality, and the worst of the worst? We try to drag others along with us to stay myopic as well, that is why we are where we are today.

Shalom!

Nice submission but you have to take it easy boss. I have a feeling that we have some things in common. When I started with Python last year September, my goal was just to do data analysis. When I 'finished' mastering the fundamentals of descriptive and inferential statistics, I said to myself; this can't be the end of it so I started learning about ML, then DL and I became obsessed with CV and NLP. After getting a solid understanding of DL, I said to myself; this can't be the end so I made some research and found out how ML/DL models can be turned to web/mobile apps so I started learning web dev (backend and fronend) till I gained proficiency in it and started building my portfolio. My next target is to start contributing to open source projects (most especially, NLP).

I think for anyone to be come a master (I mean not working for someone your entire life) in this tech space, one has to be curious and be innovative at the same time.

For anyone that wants to go deep into DS, don't listen to anyone telling you to focus on some aspects and ignore some. Get a standard learning path (check Analytics Vidhiya) and stay focused while you climb the ladder.

Between, could you please quote me with the link to that Edureka video on excel?. I don't really like their courses but let me try this one since you found it interesting. Thanks!

KlausMichaelson:



A very tricky one for me. Tricky but simple on. the long run. I really appreciate this Sir. It got me wondering if there are some Python basics I still haven't come across . Below is the code and I hope it's correct.


starting_day = int(input('Sunday through Saturday is 0 through 6, so what Day number are you leaving? '))

length_of_day = int(input('How many days will you be away? '))

Total_days = starting_day + length_of_day

Returning_day = Total_days % 7

print(f'You will return on Day {Returning_day} ')

Nice work! Keep it up. It works perfectly....

It is normal to feel like that sometimes but with practice, it gets better. Knowing only the basics of programming will not help you to overcome every single challenge you encounter, one has to be able to think algorithmically to be able to solve problems. One way to go about this is to first understand the problem you are trying to solve. After having a good understanding of the problem, translating it into code will be very simple.