Создан заказ №6324449
9 марта 2021
Доделать работу бакалавра по машинному обучению (English)
Как заказчик описал требования к работе:
Внимание - на английском языке! Правка документа на латексе по машинному обучению (в принципе не сложная, готовая, надо только доделать. времени куча)
Следует (комментарии от профессора на английском)
Here are my comments:
the abstract is still not what it should be. Try the following structur
e: tell what natural language processing is, why is it interesting; tell that you focus on an application of NLP in analyzing news text to predict the evolution of currency exchange rates; tell how you approach solving the problem; tell how successful you were with it.
Grammar comment: you seem to systematically miss the articles “a”, “the” throughout your text. The language checkers will surely catch this problem and fail you for it.
Introduction. Try to make a more verbose introduction of what NLP is, success stories, potential applications, mathematical methods used in the field, computational libraries. Then tell about your topic: why is it interesting/impactful, is data available, are such solutions already deployed in practice/companies. Only then about what you did, basically your current text. Even in this part on your contribution, tell about where the data came from, tell about the type of ML you used, tell about the structure of the rest of the thesis.
Section 3 is very short, it cannot stay like this. It is better integrated into the section with the implementation.
There should be a section on the data: where is it from, what does it contain, what is its structure, how do you curate it for your project, how do you read it into a suitable format, do you need to “digitize” it, do you need to standardize or normalize it. This can be in the place of sections 4 and 5.
There should be a section on “Artificial neural networks” where you explain the model: layers, inter-connections, activation functions, output, error functions, optimization, forward computation and back-propagation. The presentation should be independent of your data and of the NLP application in general, just focusing on what neural networks are. Some of these are already in section 6. The section can also include a discussion on the implementation of NN in Keras, as you have now in section 6. The material in section 7 would also feed into this new section.
The drawings you have in section 6 should be integrated in latex using the “Figure” environment, so they are numbered and have captions. Also, they seem a little sketchy, try some vectorial format for these pictures.
There should be a section on your model: how many layers, size of the layers, number of parameters, data size, data split into train/validate/test, number of epochs, accuracy over train/validate/test. This should replace your section 8.
In Conclusions recall why NLP is of interest, recall that NN is a powerful ML technique, recall your model and its performance. The discuss whether the accuracy you got is “enough” in a commercial setting, what could be done to improve it. Also comment on the potential of NN+NLP in practice in the future.
Insert proper references and use them throughout your text. It will bring a major improvement to your text if you read standard textbooks before you write your own version of presentation of NN and NLP.
подробнее
Заказчик
заплатил
заплатил
100 ₽
Заказчик не использовал рассрочку
Гарантия сервиса
Автор24
Автор24
20 дней
Заказчик принял работу без использования гарантии
10 марта 2021
Заказ завершен, заказчик получил финальный файл с работой
5
Доделать работу бакалавра по машинному обучению (English).docx
2021-03-13 11:04
Последний отзыв студента о бирже Автор24
Общая оценка
5
Положительно
Очень благодарна автору за проделанную работу, все пожелания учтены, заказ выполнен отлично.