The F-Score Test examines 9 tests in three areas: profitability, capital structure, and operating efficiency. Each test is given either 0 points for a fail or 1 point for a pass. All 9 tests can be derived from company financial statements.
Piotroski found that companies with 8 or 9 total points greatly improved the probability that the company had the ability to overcome challenges and eventually grow its stock price. Here are the 9 tests in the 3 areas:
Profitability (4 points)
Return on Assets
Return on assets (ROA) is net income before extraordinary items divided by total assets. If ROA is positive award 1 point; if negative, no points.
Cash Flow from Operations
Cash Flow From Operations is the cash inflows and outflows of a company’s core business operations. One point is awarded for positive cash flow from operations and none if it is negative.
Direction of Return on Assets
Did the return on assets increase this year over last year? Is so, award 1 point; if not, no points.
Accrual Accounting Check
Does the cash from operations exceed net income before extraordinary items? If yes, award 1 point, if not, no points.
Capital Structure (3 points)
Direction of Leverage
Did the long term debt to total assets ratio increase or decrease from the year before? If leverage decreased award 1 point; if it increased, no points.
Direction of Liquidity
Did the current ratio (current assets divided by current liabilities) increase or decrease from the prior year? If it increased, award 1 point; if it decreased, no points.
Issue Stock
When a company issues stock it dilutes the present shareholders stake and may indicate the company is unable to raise sufficient capital from operations. If the company has not issued stock in the past year award 1 point; if it issued stock, 0 points.
Operating Efficiency (2 points)
Direction of Margin
If the gross margin ratio has increased (year over year) that is a positive sign, award 1 point. Falling margins would be a warning sign, no points awarded.
Direction of Asset Turnover
An increase in asset turnover (sales divided by assets at the beginning of the year) from one year to another would indicate greater efficiency. Award 1 point for a higher asset turnover ratio than the previous year; 0 points if it has decreased.
Auto/None : This will simply take all the features which make sense in every tree.Here we simply do not put any restrictions on the individual tree.
sqrt : This option will take square root of the total number of features in individual run. For instance, if the total number of variables are 100, we can only take 10 of them in individual tree.”log2″ is another similar type of option for max_features.
0.2 : This option allows the random forest to take 20% of variables in individual run. We can assign and value in a format “0.x” where we want x% of features to be considered.
How does “max_features” impact performance and speed?
Increasing max_features generally improves the performance of the model as at each node now we have a higher number of options to be considered. However, this is not necessarily true as this decreases the diversity of individual tree which is the USP of random forest. But, for sure, you decrease the speed of algorithm by increasing the max_features. Hence, you need to strike the right balance and choose the optimal max_features.
1.b. n_estimators :
This is the number of trees you want to build before taking the maximum voting or averages of predictions. Higher number of trees give you better performance but makes your code slower. You should choose as high value as your processor can handle because this makes your predictions stronger and more stable.
1.c. min_sample_leaf :
If you have built a decision tree before, you can appreciate the importance of minimum sample leaf size. Leaf is the end node of a decision tree. A smaller leaf makes the model more prone to capturing noise in train data. Generally I prefer a minimum leaf size of more than 50. However, you should try multiple leaf sizes to find the most optimum for your use case.
2. Features which will make the model training easier
There are a few attributes which have a direct impact on model training speed. Following are the key parameters which you can tune for model speed :
2.a. n_jobs :
This parameter tells the engine how many processors is it allowed to use. A value of “-1” means there is no restriction whereas a value of “1” means it can only use one processor. Here is a simple experiment you can do with Python to check this metric :
model = RandomForestRegressor(n_estimator = 100,oob_score = TRUE,n_jobs = -1,random_state =1)
model.fit(X,y)
Output ———- 1 loop best of 3 : 1.1 sec per loop
“%timeit” is an awsum function which runs a function multiple times and gives the fastest loop run time. This comes out very handy while scalling up a particular function from prototype to final dataset.
2.b. random_state :
This parameter makes a solution easy to replicate. A definite value of random_state will always produce same results if given with same parameters and training data. I have personally found an ensemble with multiple models of different random states and all optimum parameters sometime performs better than individual random state.
2.c. oob_score :
This is a random forest cross validation method. It is very similar to leave one out validation technique, however, this is so much faster. This method simply tags every observation used in different tress. And then it finds out a maximum vote score for every observation based on only trees which did not use this particular observation to train itself.
Here is a single example of using all these parameters in a single function :
We have referred to Titanic case study in many of our previous articles. Let’s try the same problem again. The objective of this case here will be to get a feel of random forest parameter tuning and not getting the right features. Try following code to build a basic model :
from sklearn.ensemble import RandomForestRegressor
This is a very simplistic model with no parameter tuning. Now let’s do some parameter tuning. As we have discussed before, we have 6 key parameters to tune. We have some grid search algorithm built in Python, which can tune all parameters automatically.But here let’s get our hands dirty to understand the mechanism better. Following code will help you tune the model for different leaf size.
Exercise : Try runing the following code and find the optimal leaf size in the comment box.
By principles of the Capital Asset Pricing Model (CAPM), the script returns the normalized minimum variance and tangency vectors. These are the optimal portfolio weights for the two efficient portfolios.
How to run the script
Fork or clone
Script is initilized with OptimalPortfolioDiversification(), adding a numpy 2d array with asset returns (Stationary) as est_return_matrix arg and a Risk free return (Often just 0).
Portfolio weights are stored in (self.normalized_min_variance_vector, self.normalized_tangency_vector)
Expected return for the next period and given portfolio is stored in (self.expected_return_of_min_var_portfolio, self.expected_return_of_tangency_portfolio)
Standard deviation (Risk) for the next period and given portfolio is stored in (self.stdOfMinVariancePortfolio, self.stdOfTangencyPortfolio)
Usage examples
Im using the script to monitor daily risk in the danish C20CAP stock index, updated every 15 minutes by scraping data from borsen.dk
class OptimalPortfolioDiversification()
"""CAPM의 원리에 따라 최적의 portfolio weights를 찾음"""
def __init__(self, est_return_matrix, risk_free_return, lagrange=False):
""생성자"""
try:
"""수익률 리턴 분석 및 할당(관찰 수, 그룹 수)"""
"""평균 수익률 matrices 생성"""
"""벡터 한개 더하기"""
"""특정 해 찾기, 만약 역변환이 불가능하면 라그랑주가 자동적으로 불려짐"""
def varCovar(self):
"""수익률 매트릭스를 토대로 variance-covariance matrix를 구함"""
def minVarPortfolio(self):
"""var-covar 매트릭스를 토대로 최소 variance 포트폴리오 weight를 계산,
역변환이 불가능 할때만 적용가능, 만약 아니라면 라그랑주"""
def tangentPortfolio(self):
"""var-covar 매트릭스를 토대로 efficient/tangent 포트폴리오 weight를 계산,
역변환이 불가능 할때만 적용가능, 만약 아니라면 라그랑주"""
def lagVarCovar(self):
"""다른 함수를 위해 매트릭스들을 prepping함"""
def minVarLagrange(self):
"""라그랑주 사용해서 최소 variance portfolio 계산"""
def tangentPortfolioLagrange(self):
"""라그랑주 사용해서 최소 variance portfolio 계산"""
By the end of the specialization, you should have progressed through learning what a network is, what makes up that network, and how to connect the various components.
You also learn about the protocols and rules that govern today's IP networks.
By the end of the specialization you will be able to connect all of the components and configure them correctly to create your own functional network.
2. What is the Network?
2-1) Making the Connections
I'm going to hopefully get you started on learning all of the complexities that go on in order for you to be online. Well, online really means that you are connected to a network.
Usually, a network that is participating in the global Internet.
So basically your device, my device, any end-user device is actually connected to another group of devices, so that between me and you there may be 40, 50,100 different paths that my message from me to you can take.
What you see on the table in front of me are some of the devices that might be included.
We have end-user devices.
In any case, all of our communications between source and destination,
you and me, are intermediary devices.
These devices determine the path that your communication is going to take on its way to me,
and mine back to you.
All of these networks are managed independently meaning that different companies and
different carriers and different providers own and manage and maintain their part of the network.
So those networks, the managers and administrators of those networks agree to share a common set of rules and a common set of protocols which is what we call rules in networking,
a common set of protocols that enable communications that originate on one company's network to be able to pass through or go to destinations that are located on another network.
Now, other participants in the global Internet are educational institutions,
governments and corporations and companies of all various sizes.
There's a lot going on between you and me in order to enable us to be online.
One of the goals of this course is to actually help you untangle some of the complexity
that is involved in communications between me and you over the global Internet.
It's also to help you be able to build and understand how small networks work.
2-2) Are You Online?
Normally, when people use the term Internet, they are not referring to the physical connections in the real world. Rather, they tend to think of it as a formless collection of connections. It is the “place” people go to find or share information.
2-3) Who Owns "The Internet"?
The Internet is not owned by any individual or group. The Internet is a worldwide collection of interconnected networks (internetwork or Internet for short), cooperating with each other to exchange information using common standards. Through telephone wires, fiber-optic cables, wireless transmissions, and satellite links, Internet users can exchange information in a variety of forms, as shown in the figure.
Everything that you access online is located somewhere on the global Internet. Social media sites, multi-player games, messaging centers that provide email, online courses – all of these Internet destinations are connected to local networks that send and receive information through the Internet.
2-4) Local Networks
ocal networks come in all sizes. They can range from simple networks consisting of two computers, to networks connecting hundreds of thousands of devices. Networks installed in small offices, or homes and home offices, are referred to as Small Office Home Office (SOHO) networks. SOHO networks enable the sharing of resources, such as printers, documents, pictures and music between a few local users.
In business, large networks can be used to advertise and sell products, order supplies, and communicate with customers. Communication over a network is usually more efficient and less expensive than traditional forms of communication, such as regular mail or long distance phone calls. Networks allow for rapid communication such as email and instant messaging, and provide consolidation and access to information stored on network servers.
Business and SOHO networks usually provide a shared connection to the Internet. The Internet is considered a "network of networks" because it is literally made up of thousands of local networks that are connected to each other.
Review the different types of networks below.
Small Home Networks
Small home networks connect a few computers to each other and the Internet.
The Small Office/Home Office
The Small Office/Home Office or SOHO network enables computers within a home office or a remote office to connect to a corporate network or access centralized, shared resources.
Medium to Large Networks
Medium to large networks, such as those used by corporations and schools, can have many locations with hundreds or thousands of interconnected computers.
World Wide Networks
The Internet is a network of networks that connects hundreds of millions of computers world-wide.
2-4) Making the Connections
The Internet connects more computing devices than just desktop and laptop computers. There are devices all around that you may interact with on a daily basis that are also connected to the Internet.
For example, people are using mobile devices more every day to communicate and accomplish daily tasks, such as checking the weather or sharing pictures.
Many of the things in your home can also be connected to the Internet so that they can be monitored and configured remotely.
There are also many connected devices found in the world outside your home that provide convenience and useful or even vital information.
How many of these devices do you use on a daily basis?
Review the information below to learn more about mobile and commonly found connected devices.
Smart phones are able to connect to the Internet from almost anywhere. Smart phones combine the functions of many different products together, such as a telephone, camera, GPS receiver, media player, and touch screen computer.
Tablets, like smart phones, also have the functionality of multiple devices. With the additional screen size, they are ideal for watching videos and reading magazines or books. With on-screen keyboards, users are able to do many of the things they used to do on their laptop computer, such as composing emails or browsing the web.
A smart watch can connect to a smart phone to provide the user with alerts and messages. Additional functions, such as heart rate monitoring and counting steps, like a pedometer, can help people who are wearing the device to track their health.
Google Glass is a wearable computer in the form of glasses with a tiny screen that displays information to the wearer in a similar fashion to the Head-Up Display (HUD) of a fighter pilot. A small touch pad on the side allows the user to navigate menus while still being able to see through the Google Glass.
Many of the items in a home, such as security systems, lighting, and climate controls, can be monitored and configured remotely using a mobile device.
Television set-top boxes connect to the Internet to provide video, audio, and games on demand. These boxes can often be configured remotely to record selected programs.
Household appliances such as refrigerators, ovens, and water heaters can be connected to the Internet. This allows the homeowner to power them on or off, monitor the status of the appliance, and also be alerted to preset conditions, such as when the temperature in the refrigerator rises above an acceptable level.
A smart TV can be connected to the Internet to access content without the need for TV service provider equipment. Also, a smart TV can allow a user to browse the web, compose email, or display video, audio, or photos stored on a computer.
Many modern cars can connect to the Internet to access maps, audio and video content, or information about a destination. They can even send a text message or email if there is an attempted theft or call for assistance in case of an accident. These cars can also connect to smart phones and tablets to display information about the different engine systems, provide maintenance alerts, or display the status of the security system.
Medical devices such as pacemakers, insulin pumps, and hospital monitors provide users or medical professionals with direct feedback or alerts when vital signs are at specific levels.
Radio frequency identification (RFIDs) tags can be placed in or on objects to track them or monitor sensors for many conditions.
Weather sensors provide temperature, humidity, wind speed, and barometric pressure data.
2-5) Practice Quiz: Lab - My Online Day
3. Transmitting Data on the Network
3-1) What is Data?
Data is its raw form is the information that you input.
Every time you hit send or share or upload, you are sending data that you created somewhere out onto the network.
Now, there are different types of data.
I'm going to talk a little bit first about volunteer data.
Volunteer data is data that you offer yourself.
You realize the data is being collected about you and you are agreeing to share it or to share it somewhere on the network.
The second type of data that you may not realize is actually being collected about you is something we call inferred data.
Inferred data is data that you generate by your activities, for example, your credit card.
Every place that you use your credit card, keeps a record and therefore, if you move about and go to different stores, things are being collected about you like your preferences, the places that you typically eat or shop and also information that helps pinpoint your location.
Now, another type of data is observed data. For example, your phone probably keeps track of your location. Most people have their location services on. That information is also being stored and is being transmitted to your carrier.
hence, the term big data out in the cloud.
3-2) What Exactly is Data?
We hear about data all of the time. Customer data, personal data, health data, census data, but what exactly is data? Perhaps the simplest definition of data is that data is a value that represents something. In the physical world, we represent data as numbers, formulas, alphabetic characters, and pictures. Think about all of the data that exists just about you. Some examples include birth records, baby pictures, school records, and health records.
Most people use networks to transmit their data in order to share with others or for long-term storage. Every time you hit “send” or “share” in an app or computer application, you are telling your device to send your data to a destination somewhere on the network. Sometimes, data is being sent by your devices and you may not even be aware that this is happening. Examples of this are when you set up an automatic backup utility, or when your device automatically searches for the router in a Wi-Fi hotspot.
Review the three different categories of personal data below.
Volunteered Data
Volunteered data is created and explicitly shared by individuals, such as social network profiles. This type of data might include video files, pictures, text or audio files.
Observed Data
Observed data is captured by recording the actions of individuals, such as location data when using cell phones.
Inferred Data
Inferred data, such as a credit score, is based on analysis of volunteered or observed data.
3-3) How is Data Stored?
When your data is stored on a digital media, the data is not stored in the format that you would think. It's not letters and characters.
What it is it's bits.
Bits is a term that means it's actually an abbreviation for binary digit.
3-4) Activity - The Mighty Bit
Did you know that computers and networks only work with binary digits, zeros and ones? It can be difficult to imagine that all of our data is stored and transmitted as a series of bits. Each bit can only have one of two possible values, 0 or 1. The term bit is an abbreviation of “binary digit” and represents the smallest piece of data. Humans interpret words and pictures, computers interpret only patterns of bits.
A bit is stored and transmitted as one of two possible discrete states. This can include two directions of magnetization, two distinct voltage or current levels, two distinct levels of light intensity, or any other physical system of two discrete states. For example, a light switch can be either On or Off; in binary representation, these states would correspond to 1 and 0 respectively.
Every input device (mouse, keyboard, voice-activated receiver) will translate human interaction into binary code for the CPU to process and store. Every output device (printer, speakers, monitors, etc.) will take binary data and translate it back into human recognizable form. Within the computer itself, all data is processed and stored as binary.
Computers use binary codes to represent and interpret letters, numbers and special characters with bits. A commonly used code is the American Standard Code for Information Interchange (ASCII). With ASCII, each character is represented by eight bits. For example:
Capital letter: A = 01000001
Number: 9 = 00111001
Special character: # = 00100011
Each group of eight bits, such as the representations of letters and numbers, is known as a byte.
3-5) Getting Bits Moving
After the data is transformed into a series of bits, it must be converted into signals that can be sent across the network media to its destination. Media refers to the physical medium on which the signals are transmitted. Examples of media are copper wire, fiber-optic cable, and electromagnetic waves through the air. A signal consists of electrical or optical patterns that are transmitted from one connected device to another. These patterns represent the digital bits (i.e. the data) and travel across the media from source to destination as either a series of pulses of electricity, pulses of light, or radio waves. Signals may be converted many times before ultimately reaching the destination, as corresponding media changes between source and destination.
There are three common methods of signal transmission used in networks:
Electrical signals - Transmission is achieved by representing data as electrical pulses on copper wire.
Optical signals - Transmission is achieved by converting the electrical signals into light pulses.
Wireless signals - Transmission is achieved by using infrared, microwave, or radio waves through the air.
In most homes and small businesses, network signals are transmitted across copper wires (cables) or Wi-Fi enabled wireless connections. Larger networks employ fiber-optic cables in order to reliably carry signals for longer distances.
The animation below is a representation of three different types signals used in networks.
4. It's All about the Speed
4-1) Throughput
In this sections, we're going to demonstrate throughput using the website speedtest.net.
So throughput is the amount of data sent and received over a connection, including any delays that might occur. This includes delays in both directions.
So let's go ahead and take a look using speedtest.net
This is the number of bits per second that is sending.
There are always a good tool to help measure the performance, the amount of data you can send and receive, not just over you network, but also through your internet service provider.
4-2) Measuring Bandwidth
Streaming a movie or playing a multi-player game requires reliable, fast connections. In order to support these “high bandwidth” applications, networks have to be capable of transmitting and receiving bits at a very high rate.
Different physical media support the transfer of bits at different speeds. The rate of data transfer is usually discussed in terms of bandwidth and throughput.
Bandwidth is the capacity of a medium to carry data. Digital bandwidth measures the amount of data that can flow from one place to another in a given amount of time. Bandwidth is typically measured in the number of bits that (theoretically) can be sent across the media in a second. Common bandwidth measurements are:
Thousands of bits per second (kb/s)
Millions of bits per second (Mb/s)
Billions of bits per second (Gb/s)
Physical media properties, current technologies, and the laws of physics all play a role in determining available bandwidth.
The table shows the commonly used units of measure for bandwidth.
4-3) Measuring Throughput
Like bandwidth, throughput is the measure of the transfer of bits across the media over a given period of time. However, due to a number of factors, throughput does not usually match the specified bandwidth. Many factors influence throughput including:
The amount of data being sent and received over the connection
The types of data being transmitted
The latency created by the number of network devices encountered between source and destination
Latency refers to the amount of time, including delays, for data to travel from one given point to another.
Throughput measurements do not take into account the validity or usefulness of the bits being transmitted and received. Many messages received through the network are not destined for specific user applications. An example would be network control messages that regulate traffic and correct errors.
In an internetwork or network with multiple segments, throughput cannot be faster than the slowest link of the path from sending device to the receiving device. Even if all or most of the segments have high bandwidth, it will only take one segment in the path with lower bandwidth to create a slowdown of the throughput of the entire network.
There are many online speed tests that can reveal the throughput of an Internet connection. The figure provides sample results from a speed test.
4-4) Practice Quiz: Activity - Identify Appropriate Definitions
5. Network Components
5-1) Clients and Servers
All computers connected to a network that participate directly in network communication are classified as hosts. Hosts can send and receive messages on the network. In modern networks, computer hosts can act as a client, a server, or both. The software installed on the computer determines which role the computer plays.
Servers are hosts that have software installed that enable them to provide information, like email or web pages, to other hosts on the network. Each service requires separate server software. For example, a host requires web server software in order to provide web services to the network. Every destination that you visit online is provided to you by a server located somewhere on a network that is connected to the global Internet.
Clients are computer hosts that have software installed that enable them to request and display the information obtained from the server. An example of client software is a web browser, such as Internet Explorer, Safari, Mozilla Firefox, or Chrome.
Review more information below on descriptions of the different types of clients and servers.
File Client and Server
The File Server stores corporate end-user files in a central location. The client devices access these files with client software such as Windows Explorer.
Web Client and Server
The Web Server runs web server software and clients use their browser software, such as Windows Internet Explorer, to access web pages on the server.
Email Client and Server
The Email Server runs email server software and clients use their mail client software, such as Microsoft Outlook, to access email on the server.
5-2) Multiple Roles in the Network
A computer with server software can provide services simultaneously to one or many clients, as shown in the figure.
Additionally, a single computer can run multiple types of server software. In a home or small business, it may be necessary for one computer to act as a file server, a web server, and an email server.
A single computer can also run multiple types of client software. There must be client software for every service required. With multiple types of client software installed, a host can connect to multiple servers at the same time. For example, a user can check email and view a web page while instant messaging and listening to Internet radio.
5-3) Practice Quiz: Activity - Match the Network Components
6. Building Blocks of a Network
6-1) Networks Infrastructure Symbols
In this video, we're going to take a look at some of the symbols.
You're going to see throughout the course to represent various types of networking equipment.
In this first screen, we see some of the examples of intermediate devices.
Some of these you may recognize from your own network, such as a router, a wireless router, a switch, and a wireless access point.
Here are some examples of end devices, such as a laptop, printer, smartphone, an IP phone.
And this is example of the symbols that we'll be using for network media, such as LAN media, Local Area Network, most commonly an Ethernet LAN.
WAN media, Wide Area Network media, that's commonly used for internet service provider communications.
Wireless media, and the Cloud which is used to represent another network or the internet.
So you'll see these symbols used throughout this course and other courses.
6-2) Network Infrastructure
The path that a message takes from its source to destination can be as simple as a single cable connecting one computer to another or as complex as a network that literally spans the globe. This network infrastructure is the platform that supports the network. It provides the stable and reliable channel over which our communications can occur.
The network infrastructure contains three categories of hardware components:
Intermediate devices
End devices
Network media
Devices and media are the physical elements, or hardware, of the network. Hardware is often the visible components of the network platform such as a laptop, PC, switch, router, wireless access point, or the cabling used to connect the devices. Occasionally, some components may not be so visible. In the case of wireless media, messages are transmitted through the air using invisible radio frequency or infrared waves.
When documenting your network, it is helpful to make a list of the network infrastructure components installed in your home network. Be sure to include the cables or wireless access points that provide your network connections.
6-3) End Devices
The network devices that people are most familiar with are called end devices, or hosts. These devices form the interface between users and the underlying communication network.
Some examples of end devices are:
Computers (work stations, laptops, file servers, web servers)
Network printers
Telephones and teleconferencing equipment
Security cameras
Mobile devices (such as smart phones, tablets, PDAs, and wireless debit/credit card readers and barcode scanners)
An end device (or host) is either the source or destination of a message transmitted over the network, as shown in the animation. In order for the intended recipient to receive the message, devices in the network must be uniquely identified. In networks, this is done by assigning each host a unique address. In most networks today, an IP address is used. When a host initiates communication, it uses the address of the destination host to specify where the message should be sent.
Watch the animation of data flowing through a network.
Seeking to continuously produce high-quality trading signals (alphas) through research and simulation to employ trading strategies focused on market inefficiencies.
