My Stories

#1. MEFASH-WORLD✂

MEFASH-WORLD is an acronym for MEN FASHION GENERATING SYSTEM. The whole concept of MEFASH-WORLD was born out of the need to solve a problem in the fashion industry. Am going to be walking you through the whole process of creating a system that can generate new efficient designs, yes, fashion designs for men.

Before we start building this system there are some basic concepts we need to know and the field of DATA SCIENCE AND ANALYTICS is a field that is vital in this development process.

WHAT IS DATA SCIENCE(DS)?

Data Science is a multi-disciplinary field that brings together or combine concepts from three(3) major fields: Computer Science, Statistics/Machine Learning, and Data Analysis to understand and extract insights from the ever-increasing amount of data. One very interesting thing is that data is in abundance today and we have moved into the era of BIG DATA. Everyday we accumulate thousands of data, for instance, everyday you generate data, yes you, are you surprised how?. Ok, from the time you wake, take your bath, eat, the food you eat, the time you leave for work/school, all the activities you carry out within the day, these are all data 😃🤓.

PARADIGMS OF DATA RESEARCH:

There are actually two paradigms of data research, namely;

1. HYPOTHESIS-DRIVEN: In this approach the problem is given or the problem is identified but the data to be used is not given. So you are meant to source the data yourself.
2. DATA-DRIVEN: Here the data is given or already at hand but no problem is identified yet. For example, a company that already have registered customer and have their data, can use the data with the goal of serving their customers better.

But for our MEFASH GENERATING SYSTEM, we will be making use of the Hypothesis-Driven approach and this is because we have identified a problem and we are to source for the dataset to be used in building our system. And the dataset is going to be images.

Which paradigm of Data Research do you think is the best? And why?
Leave your answers for me in the comment session below 👇👇.Bye 🥰

#2. MEFASH GENERATING SYSTEM: ALGORITHM CLASSIFICATION METHOD

Hello guys, welcome back to another session on the how to create a FASHION GENERATING SYSTEM. A quick flashback to my first post (INTRODUCTION TO MEFASH-WORLD: Here ),I said we will be creating a system that can generate new efficient designs, yes, fashion designs for men. And yesterday I stated that we will be using the HYPOTHESIS-DRIVEN paradigm of Data Research.
We will also be making use of Machine Learning algorithms too but before that we need to know the two main types of machine learning algorithm, they are four actually.

CLASSIFICATIONS OF MACHINE LEARNING ALGORITHMS

1. SUPERVISED LEARNING: This is when algorithms are provided with training data and correct answers. All the materials are Labeled to tell the machine the corresponding value to make it predict the correct value. The task of the Machine Learning algorithm is to learn based on the training data and apply the knowledge to real data. For example, a Machine Learning algorithm to identify Happy and Unhappy pictures, You will have to feed or train your model with photos of you in your happy and unhappy mood, and then later your model will be able to identify new happy and Unhappy photos of you 😉. Hence this involves manual labeling or classification of data.
2. UNSUPERVISED LEARNING: These machine learning algorithms do not have a training dataset and no material is labeled. They are presented with some data about the real world and have to learn from that data on their own and the machine classifies materials itself by detecting the characteristics of the data. Unsupervised learning algorithms are mostly focused on finding hidden patterns in data. For example, a Machine Learning algorithm to identify if an image is of a dog or a cat. the machine will have to decide which of the picture(let's say from 100 pictures) from the data is a dog or a cat, by itself, and also classify them at the same time. Hence this does not involve manual labeling or classification of data 🤓.
3. SEMI-SUPERVISED LEARNING: Here only a few of the datasets are classified or labeled, serving as a hint to the computer. The computer finds features from the already labeled data and then classifies other data accordingly. For example, 10 out of 100 images are labeled manually and the computer labels the rest.😃
4. REINFORCEMENT LEARNING: Here the algorithm learns based on external feedback given by the environment or a thinking entity. This is applied in a machine learning algorithm that learns how to play games against an opponent. The moves that lead to victories are learned and repeated while moves that lead to failures are avoided. So here the computer is learning from experience.😁

But we are going to be making use of the UNSUPERVISED LEARNING approach for MEFASH GENERATING SYSTEM.
We will be stopping here for today, I hope you gained a thing or two from this session.
You can leave your questions about anything you are not clear about or comment on in the comment section below 👇. Bye🥰

#3. MEFASH GENERATING SYSTEM: DATA IDENTIFICATION

Hello guys 🤓 👋, I warmly welcome you to another session on how to build a FASHION GENERATING SYSTEM, in this case, MEFASH GENERATING SYSTEM 😁. Sit tight as today on this episode we will be looking at how to identify our data, that is, the kind of data that will make up our dataset and how to prepare them😊.

DATA IDENTIFICATION

In order for you to Identify the data to be used, you need to first identify the problem, in this case, our problem is how to create new fashion styles and we will be using machine learning to solve this problem. So if we are solving the problem of how to use machine learning to come up with a model that can generate new styles for customers, what do you think our dataset will look like? Now there are five types of data, that is, five forms in which our data can appear, namely:

1. STRUCTURED DATA: Data is said to be structured data if it can fit into a table, spreadsheet, or relational database. These are data that have predefined structures.
2. UNSTRUCTURED DATA: This is the opposite of structured data, they are data with no predefined structure, they come in any size or form and cannot be easily stored in a table. Examples: audio, video, images.
3. QUANTITATIVE DATA: Are mainly numerical data. Example: weight, height.
4. CATEGORICAL DATA: These are data that can be divided into groups. Example: gender, sex, tribe.
5. BIG DATA: These are very large or massive datasets and cannot fit into the memory of a single machine. They are characterized by three VS (variety, volume, velocity).

But for our MEFASH GENERATING SYSTEM, we will be making use of the Hypothesis-Driven approach and this is because we have identified a problem and we are to source for the dataset to be used in building our system. And the dataset is going to be images.To answer the question I asked earlier, we are going to be making use of the UNSTRUCTURED DATA. This is because our dataset is going to be made up of images of previous fashion styles which we will use to feed our model and train the model. This is getting interesting 😊🤓.
OK, that will be all for today guys, see you tomorrow for the next session 😉👍.

Which paradigm of Data Research do you think is the best? And why?
You can leave your questions about anything you are not clear about or comment on in the comment section below 👇. Bye🥰