Dear all,
Thankyou for most of you completing the assignment 2 quiz. If you have not completed the quiz, please do this as soon as possible so you can be well prepared to finish assignment 2. As general feedback I can say that most of you have a good hypothesis for your project. However, some of you still have issues that are likely to cause serious problems or lead you to fail the project because you do not meet learning outcome A. I can summarise these issues in relation to types of hypothesis given as follows.Classification And Types Of Burns
Probably good, but could be worded a bit better
Most of your answers fit this category. For example.
The random forests algorithm can effectively estimate the death toll of a traffic accident.
The alternating direction method of multipliers (ADMM) can be applied on recovering a blurry image corrupted by gaussian noise.
Convolutional Neural Network algorithm can classify mushroom types well based on mushroom images.
Video games especially some games containing violent elements lead to some potential
mental diseases among teenagers
The process of human beings from not understanding painting to understanding painting genres can be replicated by Deep Learning.
Assume that neural network algorithm is superior to linear regression and random forest algorithm in predicting stock price
The recommendation of singular value matrix decomposition in evd is better than the recommendation algorithm based on music similarity analysis, and it can analyze the data of the found data set offline and recommend suitable music to users.
The hypothesis of this project would be the data transfer procedure of no data center cloud model could be more efficient than the traditional client and server cloud model.
Mostly these types of hypothesis will have the potential to lead to an experiment that generates knowledge in the area of CS or engineering. These demonstrate a good understanding of how to frame proper academic research.
Probably OK, but is phrased as a question rather than a statement.
With some work you should be able to fix your hypothesis. For example.
If method based on generative adversarial networks (GAN) does better than other mothods in image super-resolution.
Whether the transfer learning would improve the LSTM in China stock price prediction
Are older people who use smartphones happier?
Can image recognition help classify multi-trash at the same time?
This paper will test three models(kNN, SVM,Random Forest) to find out which model is the most efficient and accurate in diabetes risk prediction.
Figure out which machine learning method including K-nearest neighbors, random forest and support vector ma- chine is more efficient to make classification for data of supermarket customers and find out different type of customers’ consumption behaviors by these methods applied. We can assume that three methods generate similar result for claddification.
Not possible to test because the language is too vague
You cant test this sort of hypothesis, or the result is meaningless. For example..
A lot of the online merchants are unsatisfied with today’s e-commerce platform’s business model with the very expensive and compulsory commission, so that they may need a new design of an e-commerce platform that can help to reduce merchants’ costs and enhance the efficiency and convenience of operating.
By improving the content-based movie recommendation algorithm, a movie recommendation system using user-based collaborative filtering will have an advantage in performance over content-based collaborative filtering.
Swarm intellectual algorithm will perform a promising optimization in diagnosing Cardiovascular diseases using machine learning.
The future of deep transfer learning in Brain Computer Interface (BCI) are promising ane excellent.
A hypothesis like this cannot be tested to generate new knowledge.
Trivial, the answer to the hypothesis is very obvious
There is no real point to this sort of hypothesis. Examples of trivial hypothesis are as follows..
The performance of different machine learning models are different in the prediction of China’s CPI(Consumer Price Index).
The hypothesis of this paper is whether it is possible to survey the investment population with different age needs to find out the different performance of their stock holdings and make predictive assumptions about stock market trends based on the use of data mining techniques. The software can provide sound analysis of market trends based on computer science techniques.
Two pictures with the same parts can be combined into one picture.
The hypothesis of this project would be the data transfer procedure of no data center cloud model could be more efficient than the traditional client and server cloud model.
Not a hypothesis, a statement to do something
Establish a credit card anti-fraud model to discover in advance that the customer’s credit card has been stolen.
Use linear programming method to optimize football team lineup.
A comparison of different regressions for the medical cost’s prediction.
Not an hypothesis, a general topic
Gamification effectiveness and their application in everyday life
E-commerce data warehouse (user behavior data collection)
User behavior data acquisition platform construction
In my research, The hypothesis I want to verify is that for a site with ANTI-CRAWLER MECHANISM, I will try to write PYTHON programs by analyzing its ANTI-CRAWLER STRATEGY and choosing appropriate methods to bypass the crawler, which further proves that we can write programs through Python to bypass the ANTI-CRAWLER MECHANISM.
Not a CS or Engineering topic – contribution is in another area
Use big data analytics to predict whether COVID-19 will affect Canada’s immigration trends in the future year.
This research will focus on the textual public opinion on Chinese social media (such as Sina Weibo) about the 2020 US election. The research question is what these public opinions think about this election. It includes two sub-questions: (1) What keywords are involved in these public opinions? (2) What is the sentiment of public opinion towards the two candidates? The Hypnosis would be that public opinion on Chinese social media was more positive about Biden than Trump.
With the increase of the number of motor vehicles, the problem of parking will lead to the deterioration of urban traffic conditions and the decline of people’s quality of life.
In my research, the hypothesis I want to study is that there is a high degree of facial similarity between couples, and facial similarity plays a certain role in the process of choosing a mate.
Remember. You need a proper testable hypothesis with an outcome related to Computer Science or Engineering to meet Learning outcome A and pass the module. So if your hypothesis isnt good then please work to revise it as soon as you can.
Kind regards,
Paul Craig
Module leader