Build me a NLP model to extract hidden costs from financial statements

Mbyllur Postuar 2 vite mё parё Paguhet në dorëzim
Mbyllur Paguhet në dorëzim

I would like to build a deep learning model using NLP that is able to recognize hidden costs in a 10-K or 10-Q financial statement, and extract the monetary value. There are about 7 different expense categories, each category has different keywords.

Here are some examples:

---

"Exploratory dry-hole costs were $12.7 million, $1.3 million, and $1.0 million for the years ended December 31, 2012, 2011, and 2010, respectively."

Keyword: "dry-hole costs", "2012"

Output: $12.7 million

---

"2012 includes the recognition of a $3,340 million impairment charge related to the carrying value of Citi's remaining 35% interest in the Morgan Stanley Smith Barney joint venture"

Keyword: "impairment charge"

Output: $3,340 million

---

"During the year ended December 31, 2017, we decided to discontinue the internal development of AMG 899, resulting in an impairment charge of $400 million for the IPR&D asset"

Keyword: "impairment charge"

Output: $400 million

---

"We incurred $146 million of pre-tax expenses in 2017 related to Hurricane Maria."

Keyword: "incurred ... expenses"

Output: $146 million

---

"In fiscal 2019, we recorded a $53 million charge related to the fair value adjustment of inventory acquired in the Blue Buffalo acquisition."

Keyword: "recorded a ... charge", "fair value adjustment"

Output: $53 million

---

This is just one category, I have about 100 examples of how they are applied across historic statements that can feed into an initial training set.

There are some problematic sentences that need to be avoided. For example:

"We made $100 million in profit this year, despite having significant restructuring expenses"

The algorithm should realise that although "restructuring expenses" exists in the sentence, the "$100 million" does not refer to it, but to something else and should be ignored.

There are also cases where multiple values are provided in a single sentence, and it needs to pick out the right year:

"Restructuring expenses were $40,000 in 2020, $30,000 in 2019 and $20,000 in 2018"

The correct value here should be $40,000.

I have experimented using spaCy and prodigy, but I am not sure on the best approach. One idea is to develop a NER model that recognizes if a keyword exists in a sentence, and then uses another model to parse the $ value from the sentence, using the year if necessary. It might be better to just use a single training model.

If you need any further details, please reach and out and I can give you more context.

Nxjerrje të Dhënash Përpunim i gjuhës natyrale (NLP) Inteligjencë Artificiale Python Machine Learning (ML)

ID Projekti: #29836406

Rreth projektit

16 propozimet Projekti në distancë Aktiv 2 vite mё parё

16 profesionistë freelancer dërguan një ofertë mesatare prej €636 për këtë punë

liveexperts123

Hi there,I'm biddin on your project "Build me a NLP model to extract hidden costs from financial statements" I have read your project description and i'm an expert in Machine learning/Python/C++/Java and Data science t Më shumë

€750 EUR për 4 ditë
(26 Përshtypje)
6.4
sajjadtaghvaeifr

Hi, I hope you are doing fine. I have almost 10 years of experience in machine learning algorithms. I can implement various types of artificial intelligence algorithms including yours with Matlab, Python and etc. I hav Më shumë

€500 EUR për 7 ditë
(19 Përshtypje)
5.1
techplusintl

Hi there, ★★★ Python / C++ / Machine Learning (ML) Expert ★★★ 10+ Years of Experience ★★★ I've read requirements and ready to create model to extract hidden costs from financial statements. We are a team of profession Më shumë

€750 EUR për 7 ditë
(15 Përshtypje)
5.6
snbhanja

Hi I have done similar extractive work in the past using deep learning. I am able to extract info correctly from the all example you shared except one wrong. I can retrain and fine tune the model. Let's discuss about Më shumë

€467 EUR për 4 ditë
(12 Përshtypje)
4.7
tecogno

Hi, We at Tecogno Solutions are a team of Passionate Data Science and Full Stack professionals having more than five years of combined experience in multiple areas including Backend, Frontend, Machine learning (ML), C Më shumë

€750 EUR për 7 ditë
(3 Përshtypje)
5.1
Sandeep2805

Thanks for your posting! I am a computer vision and machine learning expert with full experiences in tensorflow, darknet, keras, pytorch, opencv and open vino, etc. I have developed lots of real time face recognition p Më shumë

€750 EUR për 7 ditë
(16 Përshtypje)
5.2
vvreddy221

Hey! I am having 4+ years of Industry Experience in Machine Learning, Deep Learning,Natural Language Processing, and Computer Vision Applications. Message me to discuss more details

€500 EUR për 7 ditë
(4 Përshtypje)
4.5
nanditapatel2021

----- Build me a NLP model to extract hidden costs from financial statements ----- Hi! I'm 5 years experienced Data Scientist with experience in NLP, Data Science, Machine Learning & Deep Learning ready to do your wor Më shumë

€500 EUR për 7 ditë
(1 përshtypje)
3.8
shabeermian

hello, I have seen that you need an experienced ML expert for a NLP model to extract hidden costs from financial statements . I am a professional ML expert with more than 5 years experience. I have carefully understo Më shumë

€750 EUR për 21 ditë
(1 përshtypje)
0.0
knightmlpraput

Guten Tag. I am the CEO and Co-founder of Knight ML. Looking for ML, AI solution for your business? Over the years we have gained expertise in ML, AI, especially in Computer vision and NLP. Our team comprises of progr Më shumë

€500 EUR për 7 ditë
(0 Përshtypje)
0.0