CSCI 62 Quiz 2 Review

Topics include: Qt , Git , Software Development Lifecycle, and Data Science and Machine Learning

QMainWindow (class)

Class that your main window class inherits from that provides a framework to connect your backend .cpp file to your frontend .ui file

QWidget (class)

Class all widgets in Qt inherit from that gives them all common methods and fields. Allows for polymorphism.

ui (field)

Pointer to a list of pointers to the widgets in your UI

Connect (method)

Used to connect a slot to a signal by taking in the widget, signal for the widget, this, and the slot function.

Slots

Functions you make to respond to a signal

Signals

A trigger that is activated when an event happens (such as &QPushButton::clicked or &QTableWidget::cellClicked)

Working file/modified file

A file that has been modified from the last commit but not yet been staged or saved

Staging file

Getting a file ready to be committed but not yet committed. Done after a file is modified.

Committed file

Changes are saved to the git branch you’re working on

git add <fileName>

Moves fileName to staging status

git commit -m “description”

Commits all changes made by files currently in staging status, leaves message of description to describe the commit

git log

Lists past commits of your code

git status

Gives status of files you have in your branch

Untracked file

File that has never been committed

.gitignore

File that holds every file to ignore when committing

git diff

Shows differences between current files and most recently committed version

git commit -a -m ”description”

Stages and commits all working files and staging files.

git branch <branchName>

Creates new branch named branchName

git checkout <branch> / git switch <branch>

Switches to specified branch

git merge -m “description” <branch>

Merges specified branch with main and includes specified description.

<<<<<<< HEAD

edits in main branch

=======

edits in branchname branch

>>>>>>> branchname

File format for a merge conflict, delete <<<<<<<HEAD, =====, and >>>>>>>mybranch and choose whichever edit you want to keep,

git branch -d <branch>

Delete specified branch

Interviews

Start with a general overview question and ask follow up questions to understand actions, frustrations, skill level, variations in users, and goals. Try to understand the issue, don’t directly ask if they want an app that does X.

Personas

Task focused biography of the user we are designing for, a generic user that acts as a representation of our desired user base. Describes current behaviour

Scenarios

Describes the interactions between the persona and the system. Both the functional behaviour of the system (what it does and not how) and the goal driven behaviour of the persona.

User Requirements

High level overview of what our system should do, obtained from scenarios

System Requirements

Detailed description of system’s functions, services, and operational constraints. Obtained from user requirements and more specific.

Minimum Viable Product

Smallest version of the system that delivers value, subset of the requirements that we implement first. Designed in a way that it can be integrated and built into a full product.

Class Diagram Arrow

Shows that a class inherits from the class the arrow is pointing towards (Checking and Savings inherit from Account)

Class Diagram Multiplicity

Shows how many objects can be associated with that relationship (number is towards the class it applies to) (ex from diagram, each account only has one customer) (#..* shows that it can be # or more)

Model-View-Controller Architecture

Has a model (backend .cpp file) and a view (frontend .ui file in qt) that are connected through a controller (mainwindow.cpp in qt).

Data Science Pipeline

Collection, processing, exploration/visualization, analysis/machine learning, insight

Data Processing

Putting the raw data you collected into a format that is good for data science. Normalizing values, removing outliers, handling null values, etc.

AI

General ability of computers to emulate human thought

Data Science

Application of computational, statistical, and visualization techniques to extract and interpret knowledge from large amounts of data.

Machine Learning

Science of getting computers to realize a task without being explicitly programmed.

Supervised Learning

Give model input/output pairs to find patterns in

Unsupervised Learning

Give inputs only, model has to find a structure within them to classify them by/predict future values for

Classification

Separate data into predefined groups. Predicting a discrete-valued quantity. Can be binary (assign it 0 or 1/ true or false) or multiclass (assign it 1, 2, 3, …, k).

Regression

Predicting a continuous variable. Provide a predicted output when given a new input (what will the weather be tomorrow or what should this house cost if it is this size?)

Training data

Data we feed the model to learn from, in supervised learning this will include both input/outputs and for unsupervised learning it will be just inputs

Test data

Data we know the true values of that we feed to the model to evaluate how accurate its predictions are compared to our true values.

Confusion Matrix

Describes performance of a classification model by showing a table of true positive, false positive, true negative, and false negative values for the test data.

False Negative

A value that is actually true that the model classifies as false

False postive

A value that is actually false that the model classifies as true.

Accuracy

Ratio of correctly predicted observations to total observations. How many predictions did the model get right?

Precision

Ratio of correct predictions for either T/F to all predictions of either T/F. (Correct T/All predicted T or correct F/all predicted F).

Recall

Ratio of correct predictions either T/F to all real T/F values. (TP/(TP+FN) or (FP/(FP+TN))

F1 Score

Weighted average of precision and recall. F1>0.9 is excellent, 0.8-0.9 is good, and 0.5-0.8 is average.

Linear separator

Divides data with a line, above the line is one class below the line is a different class.

Decision Tree

More granular method of classifying data. Looks at one variable at a time and asks a question to classify the instance.