Topics Before the Midterm
What is Nominal data, Ordinal Data, Discrete Data, and Continuous Data
Ordinal Data: Groups variables into ordered categories
Discrete Data: Data that includes whole number figures (Number of people, number of doors)
Continous Data: Data can take any value (including decimals) (Height, length, temperature)
What is a Key informant interview?
What is a focus group interview?
Qualitative data collection method to gather in depth information from individuals who have specialized knowledge or insights about a particular topic and or a particular group of people
Participatory and qualitative research method involving a small diverse group of people who are part of a target population
What is supervised learning
What is unsupervised learning
Supervised Learning: A type of machine learning in which an algorithm is trained on a labeled dataset, where the outputs are already. known
Unsupervised Learning: Methods attempt to learn some kind of structure underlying the data . It does not require labeled data discovers patterns in the data without any prior knowledge
What are the Steps to Building a Dashboard?
1.Understand the Context
2.Define the objective
3.Identify your audience
4.Familiarize yourself with your data
5.Draw your dashboard on paper
6.How to sell your dashboard in an organization
What is the Food Consumption Score? What is it a proxy indicator to measure?
Composite score based on dietary diversity, food frequency and relative nutritional importance
It is calculated using the frequency of consumption of diff food groups
What does ETL mean?
Extract, Transform, and Load
Extract: Retrieves and verifies data from various sources
Transform: Processes and Organizes extracted data so it is usable
Load: Moves transformed data to a data repository
What is a PII
Any type of information relative to a physical person that can lead to its identification
What is K-Means Clustering ?
An unsupervised learning algorithm used for clustering similar data points into K groups. Doesnt require labeled data
What is a Convolutional Neural Network
Specialized type of NN where each layer consists on a set of filters that transform the input data so that the NN can learn the features of the data it is being fed
particularly effective for image recognition
What is the HDD indicator
Dietary Diversity of different food groups consumed over a given reference period.
What is an API?
Application Programming Interface
Software intermediary that allows two applications to talk to each other
What are Hallucinations, Biased and unfair models, (Risks of LLM)
Hallucinations: Incorrect information given as true
Biased and unfair models: AI models can inherit or amplify biases present in their training data leading to unfair or discriminatory outcomes
Is logistic regression, deep learning and CNNS types of supervised or unsupervised learning
Supervised
What is a Box plot helpful for?
Helpful for Summarizing multiple distributions by showing median
What is the MDD-W indicator? How is it measured?
It is a dichotomous indicator of whether or not women 15-49 years of age have consumed 5 out of 10 defined food groups the previous day or night
What is Generative Artificial Intelligence
AI models trained on very large volumes of data that are able to understand context and generate outputs based on a natural language input prompt
What is Geospatial data?
Data point that is geo-referenced
What is the Accuracy equation?
What is the Precision equation?
What is the Recall equation?
Accuracy = Correct Predictions/Total Predictions
Precision = True Positives/True Positives + False Positives
Recall = True Positives/True Positives+ False Negatives
What are recurrent neural networks
Specialized type of NN that has a recurrent layer that allows information to be looped back into the network allowing the network to maintain a memory of previous inputs
Specifically designed to deal with inputs that have a time dependency
What does the HFIAS capture
The HFIAS was designed to capture house hold behaviors signifying insufficient quality and quantity of food as well as anxiety and uncertainty over household access or food supply
What are LLMs
AI Models trained on large volumes of text, capable of understanding and generating high level of coherence and relevance
In terms of data protection what should you think about when considering proportionality?
Is the data being collected strictly for the stated purposes?
Are there measures in place to prevent the use of data for purposes other than those intially agreed upon?
Explain how deep learning works
each neuron computes a linear combination of some weights it has learned of the input it receives from the previous neuron. once the linear combination has been computed the neuron puts the output through an activation function that is used to generate a prediction or a specific action
What is the goal of Classification
To build a model that can assign input to one of several predefined categories.
The most common classification algorithms include decision trees, logistic regression, Naive Bayes
How is Food Consumption Score Measured?
Food Items are grouped into 8 food groups. And household supposed to respond the frequency of diff food groups consumed by a the household 7 days before the survey
the group is then assigned a weight based on nutrient content
, Those values summed and that is how you get the FCS score