Thứ Tư, 16 tháng 8, 2017

Youtube daily Aug 16 2017

These Barricade Door Storage Hangers are for those of you that have a 2007 and up JK and

are looking for a very safe and secure way of storing your doors while you have them

removed from your Jeep.

Now, storing the doors has always been a struggle, of course, they're big, they're heavy, they're

painted.

You don't want to scratch them, ding them, damage them in any other way.

So, there have been a lot of different solutions in the past that I've seen.

There are some carts that you can put the doors on to make them easy to roll around

on the floor, but they do take up a good bit of space and some other very inventive designs

that do take a lot of manufacturing and design.

This is going to be a very, very simple solution that as long as you have enough wall space

for your doors to go on, is going to be one of the best, safest, easiest to use, and most

secure way of storing your doors when they're not attached to your Jeep.

And for the price, I do think you're getting a really good value.

These door hangers are sealed, covered in a textured black powder coated finish, and

as you can see, they do have slotted mounting holes which makes this a very easy install.

You, of course, gonna wanna screw these directly into the studs of your wall, but you can do

that and still have some adjustability to get everything lined up.

Because this isn't just a one piece door hanger, this doesn't just have a spot to put the door

on hinge side, and then let it flap open or need to use a bungee cord or a strap to keep

it in place.

This also includes this latch plate that will allow you to close the door, keep it nice

and secure so it's not going to swing open, and damage or get in the way of anything else.

As I said before, I do think that these are really well built, and I think they're really

well thought out that they even have a small rubber bushing on the inside of these hinge

pins here, and that's tapered which will make it even that much easier to lift your doors

on and off the hangers.

So, this install is going to be a one out of three wrench installation.

It shouldn't take you more than a half hour to complete.

The first and most important step is going to be finding the studs on your wall.

You're gonna wanna screw these directly into the studs, because you do have a good bit

of weight here.

These do come with the lag bolts that you're going to need.

So, find your studs, use a stud finder, whatever method you like to use.

Make sure you have that marked, go ahead and hold this up on the wall, mark out where you

want to install those lag bolts, drill yourself a pilot, then install the lags.

Finally, you can set the doors in place.

Use the door itself as a template to figure out where to mount your latch plate.

Simply swing the door close to the wall, see where you need to mount this to get the latch

mechanism to match up to this latch plate.

And as I said, you do have a good bit of variability here.

That is a very large slotted hole, in that plate, to allow for a lot of adjustment.

This door hanging kit will run you right around $100, and I do think that's a very fair price.

Unless you have a welder and have some steel laying around, you're not going to be able

to easily replicate something that is as strong and well built as this is.

So, I do think the $100 price tag is more than fair.

So, if you're looking for a very secure and safe way to store your doors when they're

not on your Jeep, I definitely recommend taking a look at this Barricade Door Hanger Kit,

for your 2007 and up JK, that you can find right here at ExtremeTerrain.com.

For more infomation >> Jeep Wrangler Barricade Door Storage Hangers (2007-2017 JK) Review & Install - Duration: 3:35.

-------------------------------------------

Top 10 People Who Woke Up During Surgery | Extreme Surgeries - Duration: 6:14.

Top 10 People Who Woke Up During Surgery.

People who wake up during surgery are usually traumatized by the experience.

Many of them feel the surgeon�s every cut and stitch.

Unfortunately, they cannot tell anyone of the agony that they are in, as they�re usually

given muscle relaxants before the surgery.

The medication paralyzes the patient.

10.

Carol Weihrer.

Carol Weihrer spent years living in pain due to a scratched cornea.

She had undergone 17 surgeries to try to fix it, none of which worked.

Her doctor told her that they could not save her eye; it would need to be removed.

Weihrer was relieved that the pain would finally be over.Carol scheduled the surgery.

The operation started successfully.

She was given anesthesia and fell asleep.

Weihrer woke up happily, thinking her pain was finally over.

Then she heard someone say, �Cut deeper.

Pull harder.� 9.

Donna Penner.

Donna Penner needed to undergo surgery on her abdomen.

She was given anesthetic, and she woke up as nurses were washing her stomach.

Penner thought that the surgery was over.

Then she heard the surgeon ask for a scalpel.Penner felt the entire surgery: She felt the surgeon

cut her open, she felt the surgeon slide medical instruments through her abdomen, and she felt

her organs being moved around.

Penner thought that she would die.

8 .Sidney L. Williams.

Sidney L. Williams needed to undergo open heart surgery, and the surgeon warned him

that he had a 50-percent chance of dying on the table.

Williams was given anesthetic, and he drifted off to sleep.

He woke up to the sound of a bow saw�his sternum was being opened.

Williams tried to tell the doctors that he was awake, but he couldn�t make a sound,

nor could he move, speak, or see.

Williams couldn�t even cry, as the drugs he was given stopped tear production.

7.

Jeannie Smith.

Jeannie Smith needed to have both of her ovaries removed after a doctor found a cyst on one

of them.

Smith was given anesthesia, and she was wheeled into surgery.

She soon woke up.

All she could see was a bright light shining in her face.

Smith was paralyzed from the medications and could not speak.

There was no way to show the doctors she was awake.Smith had to endure every second of

the 45-minute operation 6 .Caroline Coote.

Caroline Coote needed to have surgery to remove her gallbladder.

She was placed under anesthesia, but the line that delivered the anesthetic began to leak.

The drugs wore off soon after her operation had begun, and she woke up.The muscle relaxants

worked perfectly; Coote could not move.

She was force to lie there and endure the pain of the operation.

Coote focused every ounce of her willpower into alerting the surgeons that she was awake.

She tried to scream, but she could not make a sound.

5.

Hector Alonso.

Hector Alonso needed to have a cataract operation on his right eye.

He was given anesthesia, and he drifted off.

However, Alonso woke up in the middle of the surgery.

He could see the surgeon operating on his eye.

Alonso told his surgeon that he was awake, and he begged the doctor to stop the surgery.

4 .Diana Todd.

Diana Todd was placed under anesthesia to undergo a hysterectomy.

She woke up and heard people talking.

Todd tried to hear what they were saying.

Instead, she felt the first cut.

She felt the blade slice her open again and again; she stopped counting after the fifth

cut.

Todd was hit with constant waves of pain.

The agony was indescribable.

She was trying to scream as loud as she could, but she couldn�t make a sound.

The anesthesia drugs had paralyzed her.

3 .Ms. Y.

Ms. Y had appendicitis and ovarian bleeding, and she needed surgery to correct her health

issues.

She fell asleep, but she woke up before the procedure was over.

Ms. Y heard people talking.

Then she felt a cut being made in her abdomen and something moving inside her.

She tried to open her eyes, but she couldn�t.

Ms. Y heard people saying, �So much blood!

Quickly take some pictures.�The subsequent investigation found that the anesthesia machine

had malfunctioned during the surgery, an error which went unnoticed for 25 minutes until

the surgeon saw Ms. Y�s abdominal muscles move.

The surgeon alerted the anesthesiologist, who increased her dosage.

2.

Alexandra Bythell.

Alexandra Bythell�s anesthetic wore off shortly after her appendix removal surgery

began.

She woke up and began to panic.

Her eyes were taped shut, and there were breathing tubes in her throat.

Bythell tried to breathe, but she couldn�t manage a single breath.

She thought that she was going to die.Bythell was paralyzed.

She overheard medical staff discussing her size and weight, and she felt someone poke

and prod her.

Bythell heard a shout for more morphine, and she drifted back to sleep.

1.

Sherman Sizemore.

Sherman Sizemore underwent exploratory surgery to find the cause of his abdominal pain.

He was given paralyzing drugs to stop him from moving during the surgery, but he was

not given any general anesthesia.

Sizemore was awake for the first 29 minutes of his operation.

He felt everything.Sixteen minutes after Sizemore�s stomach was sliced open, the anesthesiologist

realized his mistake.

He gave Sizemore anesthesia and an amnesia-inducing agent.

The anesthesiologist did not tell Sizemore or the other doctors about his error.

For more infomation >> Top 10 People Who Woke Up During Surgery | Extreme Surgeries - Duration: 6:14.

-------------------------------------------

Nếu Có Duyên nghe được Kinh này Ph.ật T.ổ luôn Phù Hộ Gia Đình Bạn - Duration: 31:27.

For more infomation >> Nếu Có Duyên nghe được Kinh này Ph.ật T.ổ luôn Phù Hộ Gia Đình Bạn - Duration: 31:27.

-------------------------------------------

Getting Started with Active Trader Pro® | Fidelity - Duration: 6:11.

Welcome to Getting Started with Active Trader Pro.

This demo will provide you with a quick overview

of how to navigate and customize the application

as well as a few quick tips designed to simplify

your overall experience.

Let's get started.

Our easy-to-use, intuitive toolbar was created

directly from customer feedback. It allows you

to access tools within the platform as well as

directly link to features available on Fidelity.com.

Simply click on the menu option to view the

options available within each selection.

For example, select Trade and Orders by clicking

on the menu option. To view order status,

select Orders.

To conduct research on Fidelity's award winning

research center, click on News and Research and

choose one of several research categories

including stocks, markets and sectors, an economic

calendar and more.

For easy, one-click access to critical trading

and monitoring functionality, view the shortcuts menu

located in the top left of the toolbar menu.

Simply click to view the Trade, Order Status, Quote,

or Notebook features.

Open and view multiple tools simultaneously and

create layouts that can be tailored to meet the

most simple to robust layout needs.

To help you get started we created a handful of

predefined layouts - each with a focus around a

trader's needs.

And since not every trader is alike, you can

tweak each predefined layout or create your own.

To create your own, select New Layout and enter

a layout name.

Then select the tools you'd like to view.

For this layout, we'll select Watch List, Filters,

and News.

Then arrange the windows on the desktop as you like.

Once done, save your changes using the layouts menu.

For additional help in creating or selecting a layout,

access the Active Trader Pro Setup interactive guide

from the Help menu.

To maximize screen real estate the platform includes

a "sidebar". Sidebars allow you to expand or collapse

a group of tools so they are easily available when

needed but out of the way when not.

For example, in My Layout, you've set up tools to

easily monitor the market. But you'd like to access

trading and the message center. You can expand

or collapse whenever you need to execute a trade or

view an alert.

Next we'll do a quick review of key interactions

within Active Trader Pro.

Each tool, such as this chart, offers a simplified

right click menu, providing you access to related

functionality and standard platform features

including export and settings.

You can also select "Float" to open the tool outside

of Active Trader Pro.

The action menus are symbol based and allow you to

launch additional tools seeded with the specific

symbol. For example, from the quote select the action

menu to trade the symbol or launch a chart.

Or from Watch list, mouse over a symbol to engage the

action menu. Again just click on the menu to access

the available actions.

For quick trading, click on the bid or ask from any

quote for example in a watch list, news, or quote

tool to seed a trade ticket with symbol and limit price.

Customize your workspace and experience

Select the settings option within the platform header

section to set global settings like color theme options,

displaying the market ticker and changing between

streaming or a 1 minute auto market data refresh

interval. Plus there are more options for the

individual tools.

For example, for trading, select tool settings

like skip preview.

Use the Alerts option to receive pop up alerts or

set up sounds for alerts by category. For example,

we'll select a sound alert for price.

Select the fields and ordering of columns in your

positions list and watch lists.

Simply click an item from the Available Columns list

on the left panel to place it onto the Selected Columns

list in the right panel. Then use the arrow button to

position the new column where you'd like it to appear

on the list.

You can also easily drag and drop columns in watch

list, position list, and order status for quick

reordering. For example, here we're moving the location

of the Volume column.

And drag and drop fields within the quote tool as well!

Here are a few additional quick tips to further

customize and optimize your experience.

Launch multiple instances of a tool like quotes directly

from the toolbar.

Expand a tool to adjust your view. For example, this

quote displays in two columns when the window is

expanded.

Link tools together to quickly update the symbol across

components. For example, here we're creating a new chart

and linking it to an existing quote by clicking the link

icon in both windows. Notice how the chart auto-populates

with data from the quote.

These are just a few of the feature and user experience

highlights available to help you get started.

We invite you to gain additional insights

with our education demo series,

attend a webinar, view our online materials and

to use and explore Active Trader Pro directly.

For more infomation >> Getting Started with Active Trader Pro® | Fidelity - Duration: 6:11.

-------------------------------------------

Miss. Black Caucus wants special session to change flag - Duration: 0:45.

For more infomation >> Miss. Black Caucus wants special session to change flag - Duration: 0:45.

-------------------------------------------

Best Animated Kids Series | FROG (मेढक) – Big Block – Episode 11 | Shemaroo Kids Hindi - Duration: 2:18.

Would you like to see my pond

Would you like to check it out

Would you like to come around

Come and see what it's about

So why don't you come and have a visit

I promise you, you don't want to miss it

I bet you're wondering

Where is it

Would you like to see my pond

This is where I like to splash

Come and try it it's a bash

This is where I like to sleep

When I'm sleeping, don't make a peep

This is where I like to hop

This is where I need to stop

I love it here, I'd never swap

Don't you love my little pond

Now at your own house I'm sure you've got a big old comfy bed

Not for me

I'd rather take a little snooze on my lilly pad instead

And I bet you've got some really cool stuff and yummy things to eat

That looks good

But everyone who tries my toasted bowl of flies thinks that it's a treat

Well almost everybody

Come and visit at my pond

I promise you that it'll be fun

I'll set you up with your own pad

Where you can lie out in the sun

You don't have to answer now

So just take some time and think

But, but wouldn't it be fun?

And there's plenty of water here to drink

It's important to be hydrated

For more infomation >> Best Animated Kids Series | FROG (मेढक) – Big Block – Episode 11 | Shemaroo Kids Hindi - Duration: 2:18.

-------------------------------------------

it be the first time you hear this - Duration: 2:24.

For more infomation >> it be the first time you hear this - Duration: 2:24.

-------------------------------------------

Starting Your SAS Developer Trial: Python Example - Duration: 7:15.

In this video, I'll show how to start the SAS Developer trial

and launch a Jupyter notebook.

In the notebook, I'll use Python to invoke SAS data management

and analytics capabilities on SAS Viya.

As a Python developer, this environment

should be very comfortable to you.

To begin, from the Getting Started:

SAS Viya Developer page, click the Get

Started with Python, R, and SAS API via Jupyter Notebook link.

On the next page, click Start My Trial Now.

The folders shown here contain program examples broken out

by language: SAS, Python, and R. The examples

are here to help you start exploring this environment.

Although uploading your own data is not

enabled in this trial environment,

you have access to sample data for a variety of examples,

including banking, sales, and movie ratings.

Note that many of these examples are available on the SAS GitHub

page.

A link to these examples is on the main SAS Developer Trial

page, as well as here in the notebook.

Let's open a notebook and look at an example that

uses the hmeq data set.

This is a banking example in which

you will use SAS determine which cases are bad credit risks.

In this program, we import the Python packages,

create a session with the SAS Cloud Analytics Services

(or CAS) server, and then load the data and explore it.

To prepare the data, we impute the missing values

and partition the data into training and validation data

sets.

Then we build several models, assess them,

and compare the results.

Notice there's a handy link to the documentation.

Refer to the documentation to help you

understand the SAS Python APIs for the CAS actions.

A CAS action is the smallest unit

of work for the CAS server.

CAS actions are analogous to Python functions.

CAS actions are organized into groups

called action sets, which are analogous to Python packages.

Let's review and submit each code block

as we go through the program.

In the first code block, we load the Python packages

that are needed, as well as assign variables

that we need for our modeling.

Next, we start a CAS session and load

the action sets that we use in this program.

You need to load an action set before you can call the CAS

actions contained within it.

Here, we load the data into CAS.

Now let's explore the data.

We have 11 numeric variables, and our target variable

is a binary variable BAD, which indicates

whether a loan is good or bad.

Here we'll look at the descriptive statistics

of the numeric variables.

Next, we look at the cardinality,

or the number of distinct values,

and build a graph of the missingness.

The graph shows that we have missing values

for every variable.

In order to use the data effectively,

for some of our algorithms, we impute the missing values

using a variety of imputation methods

for different variables.

Now that we have a complete data set with no missing values,

we partition the data into training and validation data

sets using stratified sampling with respect

to our target variable.

In this table, we can see that we've

divided the data into a 70/30 split

within each level of the target variable BAD.

Next, we build a decision tree model.

We also score the current data now using this model

by calling the score action for the decision tree.

Because we've saved the model as "tree model" here,

we could score new observations at a later point

using this model as well.

We also run a SAS program to add columns to the data table

for predicted probabilities of each event, which

will be used for comparing with other models

later in this program.

We submit similar code for forest,

gradient boosting machine, and neural network models.

Note that you can edit the code to configure the modeling

algorithm options as desired.

For example, let's change the number of hidden neurons

here and run this code block.

Now we assess the models using the assess action.

We also define a Python function to make it easier

to assess all four of the models with the same code.

Having assessed all four models, we

can use the assessment statistics

to create a receiver operator characteristic, or ROC

plot and a Lift plot.

They're our standard graphs for visually depicting the accuracy

and effectiveness of a model.

First, we print the area under the ROC curve for each model.

The higher the area under the curve the better,

so we can see that the Forest model outperformed

the other three.

And now, let's draw the ROC curves and Lift plots.

The ROC plot also reflects that the Forest model is the best

with the highest curve.

On the Lift chart, we see that the Forest model

has a lift as high as 5 and at the first decile

the cumulative lift is just over 3.5.

It is also fairly consistent across

the different algorithms.

When you're done making calls to CAS actions in your session,

it's a best practice to close your session.

In a non-trial environment, closing the session

releases resources for others to use.

Here in the trial environment, you have access to the system

only for the duration of your session

and no information is saved on the system.

When you come back, you start with a fresh session.

You do have the option to download notebooks for use

at a later time.

In this video, we performed some basic modeling tasks

using Python.

We encourage you to expand upon the examples that

are provided in the trial and explore the environment.

For more information, please visit us at developer.sas.com.

For more infomation >> Starting Your SAS Developer Trial: Python Example - Duration: 7:15.

-------------------------------------------

The Bulgarian Special Forces operative - #WeAreNATO - Duration: 1:01.

It's not easy to risk your life for others.

You have to love the job.

We feel that we are protecting Bulgaria.

Bulgaria is our second mother.

We are airbornes.

We jump, we have Special Operations.

That is our blood, our heart and we get up every morning with that thought.

Speed, aggression and surprise.

When you have these three points, you have a successful operation.

We learn from our partners like the American army.

They are a great help for us.

We get the best of our tactics and they get the best of theirs and mix it up.

The more you train, the better you become.

We want to be the best always.

We train for that.

We love that.

For more infomation >> The Bulgarian Special Forces operative - #WeAreNATO - Duration: 1:01.

-------------------------------------------

Children of Catholic priests live with secrets and sorrow: Jim Graham - Duration: 3:41.

All along the way I had some mysteries in my life.

I think kids knew when I was a teenager that I might not be the son of John Graham.

I asked my mother a couple times, "I don't look like anyone else in the family.

Why am I named James? There's no James in the Graham lineage."

And she just said, "I like the name James."

It didn't add up to me. But now it adds up.

My wife and I went out for a nice dinner.

I said, "My uncle, he hasn't been nice to me since my father died.

There is just something wrong."

So she said, "I'll tell you why."

She said, "John Graham wasn't your father. Your father was a Catholic priest."

I guess it's like a flash, if you lose your life and you see your whole life.

All those things just added up in seconds.

The mystery was solved about the way I was treated my whole life.

Then I was relieved at the same time that

I didn't have the genes of John Graham.

Unfortunately he never said anything nice to me.

He had no interest in anything that I did.

One time I was playing pickup football.

I was probably about 10 years of age.

I got hit. I broke my arm in two places.

John Graham came about 30 minutes later,

and he was very upset with me that I took him away from his business.

Never asked me how I felt.

He just jumped all over me, because he lost an hour of work by having to come there and sign a document.

I got some medical documents of my father.

There was a little note there — in case of emergency

who to call as next of kin.

And this nun's name appeared. So I said, she must be pretty close.

She finally said,

"My order, the mission statement is to tell the truth, and that's what I'm going to do today."

So she said, "I met your father back in the 50s.

He just had a bladder operation that was going south

and thought he may not survive, so he said,

"I have a journal in my desk; here's a key. I want you to take that and destroy it.

You can read it first, but I want you to destroy it."

And she said, "That's when I heard about you."

She told me everything she knew about my father.

The Grahams wouldn't tell me anything, and here's a stranger.

She looked at me the whole time.

She said I looked just like him.

I've written the Superior General twice, and I received a response.

And the very short response is that, "we basically don't have

that say Father Sullivan fathered a child."

If one day they would officially say I am Father Sullivan's son,

that would really be what I'd like them to say.

For more infomation >> Children of Catholic priests live with secrets and sorrow: Jim Graham - Duration: 3:41.

-------------------------------------------

Children of Catholic priests live with secrets and sorrow - Duration: 1:18.

All along the way I had some mysteries in my life.

There was just something wrong.

I was blaming myself.

By him having a child, I was his mistake.

There is nothing in church laws, rules, and regulations

that specifically mention what do you do when a priest has fathered a child.

Your father has to reject you in the name of God.

How could a calling from God be wrong?

I would find a very dark place, and I would just start to cut myself.

And every cut was the years of pain with him.

Every time there's a scandal, at least in recent years, it's always treated as if it's an isolated incident.

I connected the dots.

And I can tell you, there's a systemic problem here.

The Catholic church caused me to lose a dad.

I will never get that back.

Never.

For more infomation >> Children of Catholic priests live with secrets and sorrow - Duration: 1:18.

-------------------------------------------

Starting Your SAS Developer Trial: SAS Example - Duration: 7:21.

In this video, I'll show how to start the SAS Developer trial

and run a program in SAS Studio to illustrate

the capabilities of SAS Viya.

To begin, from the Getting Started:

SAS Viya Developer page, click Get

started with SAS Studio for the Data Scientist.

On the next page, click Start My Trial Now.

We are now in SAS Studio.

For information on using SAS Studio,

see the links on the SAS Developer trial page.

Although uploading your own data sets in this trial environment

is not enabled, you have access to sample data

for a variety of examples, including banking, sales,

and movie ratings.

Note that many of these examples are available on the SAS GitHub

page.

A link to these examples is on the main SAS Developer Trial

page.

Let's open a program and look at an example that

uses the hmeq data set.

This is a banking example in which

you use SAS to determine which cases are bad credit risks.

In this program, we create a session with the SAS Cloud

Analytics Services (or CAS) server, and then load the data

and explore it.

To prepare the data, we impute the missing values

and partition the data into training and validation

and test data sets.

Then we build several models, assess them,

and compare the results.

Notice there's a handy link to the help center documentation.

Let's review and submit the code blocks

as we go through the program.

In the first code block, we assign macro variables

to make the code easier to follow.

Here, we load the data into CAS.

Now let's explore the data.

We have 11 numeric variables, and our target variable

is the binary variable BAD, which indicates

whether a loan is good or bad.

Here we can look at descriptive statistics

of the numeric variables.

The graph shows that we have missing values

for every variable.

In order to use the data effectively,

for some of our algorithms, we impute the missing values

using a variety of imputation methods

for different variables.

We submit this code and see the output here.

Now that we have a complete data set with no missing values,

we partition the data into training, validation,

and test data sets using stratified sampling

with respect to our target variable.

In this table, we can see that we've

divided the data into a 60/30/10 split

within each level of the variable BAD.

Next, we build a decision tree model.

We'll score the data here and also save the model

so that we can use it for future use

in assessing new observations or new data that come in.

We submit similar code for forest, gradient boosting,

and neural network models.

Note that we can edit the code and configure the modeling

algorithm options as desired.

For example, let's change the number of hidden neurons

here and rerun this code block.

Now we assess the models.

We'll define a SAS macro to make it easier

to assess all four models with the same code.

Having assessed all four models, we

can use the assessment statistics

to create a receiver operating characteristic, or ROC, plot

and a cumulative lift plot.

These are standard graphs for visually depicting the accuracy

and effectiveness of a model.

And now, let's draw the ROC curve and cumulative lift

plots.

The ROC plot reflects that the gradient boosting model

is the best with the highest curve.

On the cumulative lift chart, we see

that the gradient boosting model has a lift starting just over 5

and, at the first decile or 20th percentile,

it has a lift just over 4.

It is also fairly consistent across

the different algorithms.

When you're done submitting code to the CAS

server in your session, it's a best practice

to close your session.

In a non-trial environment, closing the session

releases resources for others to use.

Here in the trial environment, you have access to the system

only for the duration of your session

and no information is saved on the system.

When you come back, you start with a fresh environment.

You do have the option to download these programs

and use them later.

In this video, we performed some basic modeling tasks in SAS.

We encourage you to expand upon the examples that

are provided in the trial and explore the environment.

For more information, please visit us at developer.sas.com.

For more infomation >> Starting Your SAS Developer Trial: SAS Example - Duration: 7:21.

-------------------------------------------

Dembele | Coutinho y Dembélé para levantar los ánimos - Duration: 4:32.

For more infomation >> Dembele | Coutinho y Dembélé para levantar los ánimos - Duration: 4:32.

-------------------------------------------

Starting Your SAS Developer Trial: R Example - Duration: 7:29.

In this video, I'll show how to start the SAS Developer trial

and launch a Jupyter notebook.

A Jupyter notebook is a web-based interface

used to submit code.

In the notebook, I'll use R to invoke SAS data management

and analytics capabilities on SAS Viya.

As an R developer, the code and the output

will look familiar, even if the interface is new to you.

To begin, from the Getting Started:

SAS Viya Developer page, click the Get

started with Python, R, and SAS API via Jupyter Notebook link.

On the next page, click Start My Trial Now.

The folders shown here contain program examples broken out

by language: SAS, Python, and R. The examples

are here to help you start exploring this environment.

Although uploading your own data sets

is not enabled in the trial environment,

you have access to sample data for a variety of examples,

including banking, sales, and movie ratings.

Note that many of these examples are available on the SAS GitHub

page.

A link to these examples is on the main SAS Developer Trial

page, as well as in the notebook.

Let's open a notebook and look at an example that

uses the hmeq data set.

This is a banking example in which

you use SAS to determine which cases are bad credit risks.

In this program, we load the R packages,

create a session with the SAS Cloud Analytics Services

(or CAS) server, and then load the data to explore it.

To prepare the data, we impute the missing values

and partition the data into training and validation data

sets.

Then we build several models, assess them,

and compare the results.

Notice there's a handy link to the documentation.

Refer to the documentation to help you

understand SAS R APIs for CAS actions.

A CAS action is the smallest unit

of work for the CAS server.

CAS actions are analogous to R functions.

CAS actions are organized into groups

called action sets, which are analogous to R packages.

Let's review and submit each block of code

as we go through the program.

In the first code block, we load the R package that is needed,

as well as assigned variables that we need for our modeling.

Next, we start a CAS session and load

the action sets that we use in this program.

You need to load the action sets before you can call the CAS

actions contained within it.

Here, we load the data into CAS.

Now let's explore the data.

We have 11 numeric variables, and our target variable

is a binary variable BAD, which indicates

whether a loan is good or bad.

Here we'll look at the first rows of our data set

and then we'll use the summary function

to look at the summary statistics for all

of our variables.

Next, we'll look at the cardinality, or the number

of distinct values, and build graphs of the missingness.

The graph shows that we have missing values

for nearly every variable.

In order to use the data effectively,

for some of our algorithms, we impute the missing values

using a variety of imputation methods

for different variables.

Now that we have a complete data set with no missing values,

we partition the data into training and validation data

sets using stratified sampling with respect

to our target variable.

In this table, we can see that we've

divided the data into a 70/30 split

within each level of the target variable BAD.

Next, we are going to map some variables to make it easier

for our code to be reused.

Next, we'll build a decision tree model

and save that model so that we can score

the data set in a future step.

We'll submit similar code for Forest, Gradient Boosting

Machines, and Neural Network models

and then we can score all of these models.

Note that you can edit the code to configure the modeling

algorithm options as desired.

For example, let's change the number of hidden neurons

and rerun this code block.

Now we will score the models that we have just built.

Using the saved score objects that

were created we'll use an R function, the lapply, in order

to concatenate all of these models

together into a single data frame.

Now we assess the models using the assess action.

We also define an R function to make it easier

to assess all four of the models with the same code.

Having assessed all four models, we

can use the assessment statistics

to create the receiver operating characteristics, or ROC plot

and we can create a misclassification chart

and confusion matrix.

And now let's draw the ROC curve.

The ROC plot also reflects that the Gradient Boosting

is the best with the highest curve.

When we're done making calls to CAS actions in your session,

it's best practice to close your session.

In a non-trial environment, closing the session

releases resources for others to use.

Here in the trial environment, you have access to the system

only for the duration of your session

and no information is saved on the system.

When you come back, you start with a fresh session.

You do have the option to download notebooks for use

at a later time.

In this video, we performed some basic modeling tasks

using R. We encourage you to expand upon the examples that

are provided in the trial and explore the environment.

For more information, please visit us at developer.sas.com.

Không có nhận xét nào:

Đăng nhận xét