Yes, DL has higher complexity compared to ML. DL uses backpropagation and many nerual layers to model highly complex systems. ML takes less time and requires less compuation compared to DL.

In Training Phase;

The time complexity of an algorithm refers to the amount of time and computational resources required to execute the algorithm, and it can be affected by a number of factors. In the context of machine learning and deep learning, the time complexity of the training phase can be influenced by the following factors:

In general, deep learning algorithms can be more computationally intensive than some traditional machine learning algorithms during the training phase. This is because deep learning algorithms often involve the use of multiple layers of artificial neural networks, which require significant amounts of computation to train. These neural networks can be particularly useful for modeling complex relationships in data, but they can also require more time and computational resources to train than some other machine learning algorithms.

In Testing Phase;

The time complexity of the testing phase in machine learning and deep learning algorithms can vary depending on a number of factors, including the specific algorithm being used and the characteristics of the dataset being tested. Some machine learning algorithms may be faster to test than others, while some deep learning algorithms may be faster to test than others.

One factor that can influence the time complexity of the testing phase is the complexity of the model being tested. Models that are more complex, such as deep neural networks with many layers and a large number of parameters, may require more time and computational resources to test than simpler models. The size of the dataset being tested can also affect the time complexity of the testing phase, as the model will need to process more data and potentially more features, which can be computationally intensive.

It is difficult to make a general statement about the time complexity of the testing phase for machine learning algorithms compared to deep learning algorithms, as it will depend on the specific circumstances. For example, a simple linear regression model may be faster to test than a deep neural network with many layers, while a deep neural network with fewer layers may be faster to test than a more complex machine learning algorithm. The time complexity of the testing phase can also be influenced by the hardware being used to run the algorithm, as well as the specific implementation of the algorithm.

DEAR Rajnish Kumar

Thanks a lot for your idea

DEAR Uditha Prabhath Bandara Herath

Thank you very much for your detailed explanation

My hardware resources and data sets are completely the same. According to the mentioned conditions, is my conclusion correct? I look forward to your reply

my conclusion is,

In general, the training phase of deep learning algorithms tends to have a higher time complexity than that of machine learning algorithms, while the time complexity of the testing phase can vary depending on the specific algorithm and dataset being used.

Deep learning algorithms often involve training large, complex neural networks on large datasets, which can be computationally intensive and time-consuming, while many machine learning algorithms are relatively simple and can be trained more quickly.

However, the time complexity of both the training and testing phases can vary significantly depending on the specific algorithm and dataset being used, and it is not always the case that one type of algorithm is faster than the other in all situations.

If you are using the same hardware resources and data sets, then the time complexity of the training and testing phases of machine learning and deep learning algorithms should be similar.

However, it's important to note that there are many factors that can impact the time complexity of an algorithm, including the specific algorithm being used, the complexity of the model being trained or tested, and the size and complexity of the data being used.

further more, It's also worth noting that the time complexity of an algorithm is only one factor to consider when deciding which algorithm to use for a particular task. Other factors to consider may include the accuracy and performance of the algorithm, the ease of implementation and use, and the general applicability of the algorithm to different types of problems.

Overall, it's important to consider all of these factors when selecting an algorithm for a particular task, rather than simply focusing on the time complexity of the algorithm.

Thank you very much and your explanation is very complete

The computational complexity of machine learning and deep learning is different in my experiments

And deep learning is more because it uses a balanced data set (the ratio of positive to negative class is also 1:1) and machine learning from an unbalanced data set (the ratio of positive to negative class is also 1:17)

Fortunately, in deep learning, despite being time-consuming, performance such as F-M, sensitivity, and accuracy have improved.

https://libraryapp.uomustansiriyah.edu.iq/ebooks.php?a=view&recid=2768

DEAR Omar Laith

Thanks a lot for your idea

DEAR Ramtin Zargari Marandi

Thanks a lot for your accurate and complete explanation

May I ask your more explanation about "techniques such as transfer learning to fine-tune pre-trained models?"Thanks in advanced