I am searching for some research topics on machine learning, something that is suitable for an undergraduate student. I would be glad if anyone can provide me with trending topics. Thanks in advance.
You need a topic on which there is some open data. It should also be a topic that interests you, since domain knowledge is invaluable. Within those two limits, there are huge number of possibilities. For example, I recently heard a great presentation by a Mixed Martial Arts fan on predicting the results of fights.
What is "trending" is not application domains but techniques. For an undergraduate project, I would suggest you stick with established techniques, but apply them to a new question.
Since you are an undergraduate student, you are on the initial steps on AI and ML. I try to first mention the general fields and then have a closer view of each and give some examples.
Generally to start with, existing applications can be divided into:
1- Text Mining and Text Classification: categorizing datasets of text and extract patterns out of them to predict the unseen samples in the dataset. For example: Spam and Non-Spam Email detection (text classification), Classification of Words and Phrases into Predefined Classes (Extracting similarity), Extraction of Keywords from Sentences and Classification of the Sentences According to Those Keywords, Prediction of Students Status/Grades According to Their Previous Records, Bank Customer's Prediction (e.g. Loan Payback)
2- Image-based applications: In image applications one must first get familiar with masks, convolution, edge and corner detection to be able to extract useful information from images and further use them for one of the below applications. For example: Image Segmentation, Keypoints Extraction, Steganography Using Image Interpolation, Shadow removal, Image Background Brushing, Coloring gray-Images with Prediction According to Image Details.
3- Machine Vision: Using machine learning-based/mathematical techniques to enable machines do specific tasks. For example: Watermarking, Face Identification from Datasets of Image with Rotation and Different Camera Angles, Criminals Identification From Surveillance Cameras (Video and series of Image), Handwriting and Personal Signature Classification, Object Detection/Recognition
4- Clustering: Categorize data/text/image without training a model but using mathematical/statistical and geometrical techniques. For example: Graph Clustering, Data Clustering, Density-based Clustering, Decision Graphs/Trees, Constrained/Unconstrained Clustering
5- Optimization:
A) Population-based optimization inspired from a natural mechanism: Black-Box Optimization, Multi/Many-Objective Optimization, Evolutionary Methods (Genetic Algorithm, Genetic Programming, Memetic Programming), Metaheuristics (e.g. PSO, ABC, SA),
B) Exact/Mathematical Models: Convex Optimization, Bi-Convex and Semi-Convex Optimization, Gradient Descent, Block Coordinate Descent, Manifold Optimization and Algebraic Models
6- Voice Classification: Voice and Speech Recognition, Signal Processing, Message Embedding, Message Extraction from Voice Encoded
7- Other Tasks: Dataset Collection/Creation, Medical Applications like Computational Biology, ...
Big Data is a recent area where you can apply a machine-learning algorithm to extract information as per your requirement depending on the data set you have chosen. Like Wikipedia which is still BIg Data, You can process it using Scala with Big Graph libraries. So you have to do text mining first and then go ahead as per your requrement.
My humble request is to identify an real world use case and try it in ML. For Example : Identification cracks on an iron rod. Pattern to identify rusted rail tracks. Identifying the quality of the Milk sample or any other grocery sample. If this higher to your scope. Try its base version so that you can enhance it in your higher studies.