Week 6 — Career Path Predictor

Hello everyone from a new week where we spend every day working on Neural Network! Of course, our project wasn’t the only reason why every day was like this, remember that we are artificial intelligence engineering students. And we love this sport :)

This week, we are starting this blog post by reporting the developments in our project. This week, we started our work with the Neural Network application, which is the method in the article we took as an example for our project. For this, we first created a simple neural network model suitable for our data. The accuracy results we received did not make us happy, because it was written in the article that the work increased from 20% accuracy values to 80%. However, we couldn’t go much further than 20%.

With the results we got from here, we made it clear what kind of project we want our project to turn into at the end. We are explaining to you the surprise we talked about last week. We are planning to make a general job prediction with our large dataset and then make a more detailed job suggestion or even suggestions with our smaller dataset for you, our beautiful users. Thus, we aim to increase our assistance to you in the decision-making process.

Let’s continue with our other works this week. While searching for suitable models for our problem, we came across a pretrained model, AdaNet. Then we did research and examined the open source code to understand how this model works. In the implementation phase, we tried to adapt one of the existing examples to our study. Unfortunately, we couldn’t adapt it because we couldn’t overcome the version inconsistencies as a result of a very long effort. Due to the problems that took up too much of our time, we did not have time to make our simple model more suitable for our problem. As the project progressed, we decided to move forward by making our Neural Network model more complex.

Letting you know that we are at the beginning of a tiring process for us, we bid you farewell here to write our Progress Report for our project. See you next week, bye!

-Melike Nur DULKADİR

-Sare Naz ERSOY