Predict with confidence : Tensorflow
April 12, 2019
Predict with confidence : Tensorflow
Score 8 out of 10
Vetted Review
Verified User
Overall Satisfaction with TensorFlow
We use TensorFlow for machine learning implementations. Primarily for predictive analysis and recommendation engines. It is being used at an organization level. Our objective is to use a large amount of publicly available data and make meaningful insights from it. It has helped us make better predictions and save costs.
We also use it for time series analysis to make predictions in the equity market. TensorFlow has been a powerful and easy to deploy tool for various algorithms.
We also use it for time series analysis to make predictions in the equity market. TensorFlow has been a powerful and easy to deploy tool for various algorithms.
- Support for many libraries and programming languages.
- Ability to use GPU and TPU - hence faster execution.
- Low effort in getting started in development, hence ease of learning.
- Graphic interface to create layers can help beginners.
- Detailed tutorials on what goes behind the scenes in each layer. Currently, the tutorials don't focus on that.
- Better support to integrate with files on the cloud.
- Ability to make better predictions.
- Increase in profit from equity investments on a consistent basis.
- Move towards digital transformation in the company and a better brand name.
Most of the machine learning platforms these days support integration with R and Python libraries. So, the use of reusable libraries is not an issue. TensorFlow performs well in cloud hosting and support for GPU/TPU. However, where it lacks compared to Azure is a graphical front-end to drag and drop layers.
Using TensorFlow
3 - Currently, we use Tensorflow to develop algorithm-based trading models. These are time series based predictive models using NSE data publicly available. It is being used by the investment modeling department of our business.
3 - Analytical thinking is a must. A good understanding of statistics, probability, and matrices. Logical thinking, project experience in at least one of the machine learning platforms/languages like Python, R, Azure, will help the use of TensorFlow.
- Predictive Analytics - algorithm based trading
Evaluating TensorFlow and Competitors
Yes - We were using R machine learning with Shinyapps. TensorFlow was easier to implement with better support for online/cloud hosting.
- Product Features
- Product Usability
- Analyst Reports
- Third-party Reviews
Product features: Tensor flow comes with the support of built-in algorithms that are easy to implement.
We would now consider a lot more tools that have been released.
TensorFlow Implementation
- Implemented in-house
Yes - We started with smaller problem statements and took them to completion. Then added more features.
Change management was minimal
TensorFlow Support
| Pros | Cons |
|---|---|
None | None |
No - not needed so far.
At times when we got stuck with some code, use of open source forums was the way to go for problem resolution. We found support from the community forum members.
Using TensorFlow
| Pros | Cons |
|---|---|
Like to use Relatively simple Easy to use Well integrated Consistent Quick to learn Convenient Feel confident using | None |
- Adding of new neural network layers in the code.
- Running the model. Especially in the newer versions where a number of epochs and other execution parameters are easy to use.
- support for Keras, Numpy, Pandas and other libraries.
- Graphical front-end to develop code.