First of all I would like to thank Kaggle team and various data donators for this challenging data-set. Starter kernel is posted is here
\n\nI have mentioned scores wherever I have found some landmarks:\npubt : public train accuracy\nprit : private train accuracy
\n\nWhat I tried\n- First tried running with all parameters included just to get a baseline score (pubt : lower 0.6x, prit : higher 0.5x).\n- Next I thought to create columns featuring differences in adjacent values. This showed some promising results also this features were having higher weightage predicted by lgbm. However, I was still having some doubts related to orientation values (pubt : lower 0.6x, prit : higher 0.5x).\n- Removed all orientation columns and reran, Values fell straight, but now it appears it has better private score (pubt : higher 0.4x, prit : lower 0.6x).\n- Later, I tried to predict robot's velocity with various factors present but had to give this due to \nunsatisfying results.\n- Meanwhile, Christoffer Karlson frequency] established Sampling frequency. This turned my attention to frequency domain and trying to use fft's here.\n- As I was still testing, Thomas Rowher shared this awesome FFT Analysis and shared his insights here.\n- Also, I was smoothing the values present in signal, with Savitzky-Golay filter rather than averaging them out. This is great to increase SNR without distorting signal too much.\n- I started to work heavily with fft's(peaks, range, power) and I found some improvements but still I felt something was missing (pubt : lower 0.7x, prit : higher 0.6x).\n- I was thinking if maybe we could have more values in every set, and Markus F had an amazing find - Links between sets. \n- Next, I tried to use more samples in every set, however I think maybe I miscalculated and had some error which resulted in lower scores. I gave up this idea.\n- Finally, I linked all the test set & training sets for fun. Thought what we can see if we linked entire test with train. \n- I found a total of 151 linked sets. In which, 104 sets contained some data from train sets. These were somewhat easy to predict or so I thought. Remaining 47 test sets were not linked to any train data. Here I put in data which I modeled with RF. And with this public train score raised to 0.96. It appeared as if I had hit gold (pubt: 0.96, prit: 0.68). I will soon publish a kernel which uses Markus links along with RF model which gives this score. Posted, check link above :)\n- However, I had a feeling this was somehow overfitting the train and public test sets. It was!!\n- I then tried to check for all the sets where different surfaces got linked. Once plotted, it was somewhat clear which were wrongly linked. Did some post-processing and reran(pubt: 0.98, prit: 0.76).
\n\nThis is the first competition I have participated strongly and it was great to learn from many people here. I would like to specially thank :
\n\nThomas Rohwer\nMarkus F\nChristoffer Karlsson\nGilles Vandewiele\nNanashi
\n\nLooking forward for more solutions from everyone :)
\n\nPS. The idea underneath this kernel was floating on forum for more than 2 days before the competition end.
", "comments": [ { "id": 515104, "author_name": "Eric R", "author_url": "", "post_date": "2019-04-12T07:43:59.403000", "content": "Congrats! Look forward to your kernel.
", "votes": 0, "replies": [ { "id": 515549, "author_name": "Rajanikant Tenguria", "author_url": "", "post_date": "2019-04-12T18:56:18.627000", "content": "HI Eric, submitted
", "votes": 0, "replies": [] } ] } ], "comment_count": 2, "messages": [ { "id": 514956, "postDate": "2019-04-12T04:07:42.223Z", "content": "First of all I would like to thank Kaggle team and various data donators for this challenging data-set. Starter kernel is posted is here
\n\nI have mentioned scores wherever I have found some landmarks:\npubt : public train accuracy\nprit : private train accuracy
\n\nWhat I tried\n- First tried running with all parameters included just to get a baseline score (pubt : lower 0.6x, prit : higher 0.5x).\n- Next I thought to create columns featuring differences in adjacent values. This showed some promising results also this features were having higher weightage predicted by lgbm. However, I was still having some doubts related to orientation values (pubt : lower 0.6x, prit : higher 0.5x).\n- Removed all orientation columns and reran, Values fell straight, but now it appears it has better private score (pubt : higher 0.4x, prit : lower 0.6x).\n- Later, I tried to predict robot's velocity with various factors present but had to give this due to \nunsatisfying results.\n- Meanwhile, Christoffer Karlson frequency] established Sampling frequency. This turned my attention to frequency domain and trying to use fft's here.\n- As I was still testing, Thomas Rowher shared this awesome FFT Analysis and shared his insights here.\n- Also, I was smoothing the values present in signal, with Savitzky-Golay filter rather than averaging them out. This is great to increase SNR without distorting signal too much.\n- I started to work heavily with fft's(peaks, range, power) and I found some improvements but still I felt something was missing (pubt : lower 0.7x, prit : higher 0.6x).\n- I was thinking if maybe we could have more values in every set, and Markus F had an amazing find - Links between sets. \n- Next, I tried to use more samples in every set, however I think maybe I miscalculated and had some error which resulted in lower scores. I gave up this idea.\n- Finally, I linked all the test set & training sets for fun. Thought what we can see if we linked entire test with train. \n- I found a total of 151 linked sets. In which, 104 sets contained some data from train sets. These were somewhat easy to predict or so I thought. Remaining 47 test sets were not linked to any train data. Here I put in data which I modeled with RF. And with this public train score raised to 0.96. It appeared as if I had hit gold (pubt: 0.96, prit: 0.68). I will soon publish a kernel which uses Markus links along with RF model which gives this score. Posted, check link above :)\n- However, I had a feeling this was somehow overfitting the train and public test sets. It was!!\n- I then tried to check for all the sets where different surfaces got linked. Once plotted, it was somewhat clear which were wrongly linked. Did some post-processing and reran(pubt: 0.98, prit: 0.76).
\n\nThis is the first competition I have participated strongly and it was great to learn from many people here. I would like to specially thank :
\n\nThomas Rohwer\nMarkus F\nChristoffer Karlsson\nGilles Vandewiele\nNanashi
\n\nLooking forward for more solutions from everyone :)
\n\nPS. The idea underneath this kernel was floating on forum for more than 2 days before the competition end.
", "rawMarkdown": "First of all I would like to thank Kaggle team and various data donators for this challenging data-set. Starter kernel is posted is [here](https://www.kaggle.com/algorrt/highest-scoring-public-kernel-starter-sol-29?scriptVersionId=12845117)\n\nI have mentioned scores wherever I have found some landmarks:\npubt : public train accuracy\nprit : private train accuracy\n\nWhat I tried\n- First tried running with all parameters included just to get a baseline score (**pubt : lower 0.6x, prit : higher 0.5x**).\n- Next I thought to create columns featuring differences in adjacent values. This showed some promising results also this features were having higher weightage predicted by lgbm. However, I was still having some doubts related to orientation values (**pubt : lower 0.6x, prit : higher 0.5x**).\n- Removed all orientation columns and reran, Values fell straight, but now it appears it has better private score (**pubt : higher 0.4x, prit : lower 0.6x**).\n- Later, I tried to predict robot's velocity with various factors present but had to give this due to \nunsatisfying results.\n- Meanwhile, [Christoffer Karlson](https://www.kaggle.com/christoffer) frequency] established [Sampling frequency](https://www.kaggle.com/christoffer/establishing-sampling-frequency). This turned my attention to frequency domain and trying to use fft's here.\n- As I was still testing, [Thomas Rowher](https://www.kaggle.com/trohwer64) shared this awesome [FFT Analysis](https://www.kaggle.com/trohwer64/simple-fourier-analysis) and shared his insights [here](https://www.kaggle.com/c/career-con-2019/discussion/87239#503272).\n- Also, I was smoothing the values present in signal, with [Savitzky-Golay](https://en.wikipedia.org/wiki/Savitzky%E2%80%93Golay_filter) filter rather than averaging them out. This is great to increase SNR without distorting signal too much.\n- I started to work heavily with fft's(peaks, range, power) and I found some improvements but still I felt something was missing (**pubt : lower 0.7x, prit : higher 0.6x**).\n- I was thinking if maybe we could have more values in every set, and [Markus F](https://www.kaggle.com/friedchips) had an amazing find - [Links between sets](https://www.kaggle.com/friedchips/the-missing-link). \n- Next, I tried to use more samples in every set, however I think maybe I miscalculated and had some error which resulted in lower scores. I gave up this idea.\n- Finally, I linked all the test set & training sets for fun. Thought what we can see if we linked entire test with train. \n- I found a total of 151 linked sets. In which, 104 sets contained some data from train sets. These were somewhat easy to predict or so I thought. Remaining 47 test sets were not linked to any train data. Here I put in data which I modeled with RF. And with this public train score raised to 0.96. It appeared as if I had hit gold (**pubt: 0.96, prit: 0.68**). I will soon publish a kernel which uses Markus links along with RF model which gives this score. Posted, check link above :)\n- However, I had a feeling this was somehow overfitting the train and public test sets. It was!!\n- I then tried to check for all the sets where different surfaces got linked. Once plotted, it was somewhat clear which were wrongly linked. Did some post-processing and reran(**pubt: 0.98, prit: 0.76**).\n\nThis is the first competition I have participated strongly and it was great to learn from many people here. I would like to specially thank : \n\n[Thomas Rohwer](https://www.kaggle.com/trohwer64)\n[Markus F](https://www.kaggle.com/friedchips)\n[Christoffer Karlsson](https://www.kaggle.com/christoffer)\n[Gilles Vandewiele](https://www.kaggle.com/group16)\n[Nanashi](https://www.kaggle.com/jesucristo)\n\nLooking forward for more solutions from everyone :)\n\nPS. The idea underneath this kernel was floating on forum for more than 2 days before the competition end. ", "votes": 3 }, { "id": 515104, "postDate": "2019-04-12T07:43:59.403Z", "content": "Congrats! Look forward to your kernel.
", "rawMarkdown": "Congrats! Look forward to your kernel.", "replies": [ { "id": 515549, "postDate": "2019-04-12T18:56:18.627Z", "content": "HI Eric, submitted
", "rawMarkdown": "HI Eric, [submitted](https://www.kaggle.com/algorrt/highest-scoring-public-kernel-starter-sol-29)" } ] } ], "complete": true, "source": "live", "fetched_at": "2026-10-19T15:45:03Z", "competition_id": 13242, "competition_slug": "career-con-2019", "team_id": 2941272, "solution_url": "/competitions/career-con-2019/writeups/surfacefinder-highest-publiclb-kernel-posted-29-so" }