Uni/machine-learning
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add initial lines of code

Browse Source
This commit is contained in:
Tim
2025-08-11 18:57:11 +02:00
parent 4746c21729
commit 906d4711f1
1 changed files with 271 additions and 106 deletions
Download Patch File Download Diff File Expand all files Collapse all files

377
notebook.ipynb
View File

@@ -6,18 +6,16 @@
"metadata": {},
"source": [
"# Machine Learning project in SoSe 2025 at HTW Saar\n",
"### Contributors\n",
" - Maximilian Kany\n",
" - Florian Speicher\n",
" - Tim Wall\n",
"## Idea\n",
"The goal of this project is getting the genre(s) of a game trough its given metadata\n",
"\n",
"#### Dataset\n",
"For our project we use a steam DataSet from kaggle. You can find it under the following url: [Kaggle.com](https://www.kaggle.com/datasets/trolukovich/steam-games-complete-dataset)"
"## Dataset\n",
"For our project we use a Steam DataSet from kaggle. You can find it under the following URL: [Kaggle.com](https://www.kaggle.com/datasets/trolukovich/steam-games-complete-dataset)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"execution_count": 19,
"id": "3116b75f",
"metadata": {},
"outputs": [
@@ -25,133 +23,300 @@
"name": "stdout",
"output_type": "stream",
"text": [
" 0 1 \\\n",
"0 url types \n",
"1 https://store.steampowered.com/app/379720/DOOM/ app \n",
"2 https://store.steampowered.com/app/578080/PLAY... app \n",
"3 https://store.steampowered.com/app/637090/BATT... app \n",
"4 https://store.steampowered.com/app/221100/DayZ/ app \n",
" url types \\\n",
"0 https://store.steampowered.com/app/379720/DOOM/ app \n",
"1 https://store.steampowered.com/app/578080/PLAY... app \n",
"2 https://store.steampowered.com/app/637090/BATT... app \n",
"3 https://store.steampowered.com/app/221100/DayZ/ app \n",
"4 https://store.steampowered.com/app/8500/EVE_On... app \n",
"\n",
" 2 \\\n",
"0 name \n",
"1 DOOM \n",
"2 PLAYERUNKNOWN'S BATTLEGROUNDS \n",
"3 BATTLETECH \n",
"4 DayZ \n",
" name \\\n",
"0 DOOM \n",
"1 PLAYERUNKNOWN'S BATTLEGROUNDS \n",
"2 BATTLETECH \n",
"3 DayZ \n",
"4 EVE Online \n",
"\n",
" 3 \\\n",
"0 desc_snippet \n",
"1 Now includes all three premium DLC packs (Unto... \n",
"2 PLAYERUNKNOWN'S BATTLEGROUNDS is a battle roya... \n",
"3 Take command of your own mercenary outfit of '... \n",
"4 The post-soviet country of Chernarus is struck... \n",
" desc_snippet \\\n",
"0 Now includes all three premium DLC packs (Unto... \n",
"1 PLAYERUNKNOWN'S BATTLEGROUNDS is a battle roya... \n",
"2 Take command of your own mercenary outfit of '... \n",
"3 The post-soviet country of Chernarus is struck... \n",
"4 EVE Online is a community-driven spaceship MMO... \n",
"\n",
" 4 \\\n",
"0 recent_reviews \n",
"1 Very Positive,(554),- 89% of the 554 user revi... \n",
"2 Mixed,(6,214),- 49% of the 6,214 user reviews ... \n",
"3 Mixed,(166),- 54% of the 166 user reviews in t... \n",
"4 Mixed,(932),- 57% of the 932 user reviews in t... \n",
" recent_reviews \\\n",
"0 Very Positive,(554),- 89% of the 554 user revi... \n",
"1 Mixed,(6,214),- 49% of the 6,214 user reviews ... \n",
"2 Mixed,(166),- 54% of the 166 user reviews in t... \n",
"3 Mixed,(932),- 57% of the 932 user reviews in t... \n",
"4 Mixed,(287),- 54% of the 287 user reviews in t... \n",
"\n",
" 5 6 \\\n",
"0 all_reviews release_date \n",
"1 Very Positive,(42,550),- 92% of the 42,550 use... May 12, 2016 \n",
"2 Mixed,(836,608),- 49% of the 836,608 user revi... Dec 21, 2017 \n",
"3 Mostly Positive,(7,030),- 71% of the 7,030 use... Apr 24, 2018 \n",
"4 Mixed,(167,115),- 61% of the 167,115 user revi... Dec 13, 2018 \n",
" all_reviews release_date \\\n",
"0 Very Positive,(42,550),- 92% of the 42,550 use... May 12, 2016 \n",
"1 Mixed,(836,608),- 49% of the 836,608 user revi... Dec 21, 2017 \n",
"2 Mostly Positive,(7,030),- 71% of the 7,030 use... Apr 24, 2018 \n",
"3 Mixed,(167,115),- 61% of the 167,115 user revi... Dec 13, 2018 \n",
"4 Mostly Positive,(11,481),- 74% of the 11,481 u... May 6, 2003 \n",
"\n",
" 7 8 \\\n",
"0 developer publisher \n",
"1 id Software Bethesda Softworks,Bethesda Softworks \n",
"2 PUBG Corporation PUBG Corporation,PUBG Corporation \n",
"3 Harebrained Schemes Paradox Interactive,Paradox Interactive \n",
"4 Bohemia Interactive Bohemia Interactive,Bohemia Interactive \n",
" developer publisher \\\n",
"0 id Software Bethesda Softworks,Bethesda Softworks \n",
"1 PUBG Corporation PUBG Corporation,PUBG Corporation \n",
"2 Harebrained Schemes Paradox Interactive,Paradox Interactive \n",
"3 Bohemia Interactive Bohemia Interactive,Bohemia Interactive \n",
"4 CCP CCP,CCP \n",
"\n",
" 9 \\\n",
"0 popular_tags \n",
"1 FPS,Gore,Action,Demons,Shooter,First-Person,Gr... \n",
"2 Survival,Shooter,Multiplayer,Battle Royale,PvP... \n",
"3 Mechs,Strategy,Turn-Based,Turn-Based Tactics,S... \n",
"4 Survival,Zombies,Open World,Multiplayer,PvP,Ma... \n",
" popular_tags \\\n",
"0 FPS,Gore,Action,Demons,Shooter,First-Person,Gr... \n",
"1 Survival,Shooter,Multiplayer,Battle Royale,PvP... \n",
"2 Mechs,Strategy,Turn-Based,Turn-Based Tactics,S... \n",
"3 Survival,Zombies,Open World,Multiplayer,PvP,Ma... \n",
"4 Space,Massively Multiplayer,Sci-fi,Sandbox,MMO... \n",
"\n",
" 10 \\\n",
"0 game_details \n",
"1 Single-player,Multi-player,Co-op,Steam Achieve... \n",
"2 Multi-player,Online Multi-Player,Stats \n",
"3 Single-player,Multi-player,Online Multi-Player... \n",
"4 Multi-player,Online Multi-Player,Steam Worksho... \n",
" game_details \\\n",
"0 Single-player,Multi-player,Co-op,Steam Achieve... \n",
"1 Multi-player,Online Multi-Player,Stats \n",
"2 Single-player,Multi-player,Online Multi-Player... \n",
"3 Multi-player,Online Multi-Player,Steam Worksho... \n",
"4 Multi-player,Online Multi-Player,MMO,Co-op,Onl... \n",
"\n",
" 11 12 \\\n",
"0 languages achievements \n",
"1 English,French,Italian,German,Spanish - Spain,... 54 \n",
"2 English,Korean,Simplified Chinese,French,Germa... 37 \n",
"3 English,French,German,Russian 128 \n",
"4 English,French,Italian,German,Spanish - Spain,... NaN \n",
" languages achievements \\\n",
"0 English,French,Italian,German,Spanish - Spain,... 54.0 \n",
"1 English,Korean,Simplified Chinese,French,Germa... 37.0 \n",
"2 English,French,German,Russian 128.0 \n",
"3 English,French,Italian,German,Spanish - Spain,... NaN \n",
"4 English,German,Russian,French NaN \n",
"\n",
" 13 \\\n",
"0 genre \n",
"1 Action \n",
"2 Action,Adventure,Massively Multiplayer \n",
"3 Action,Adventure,Strategy \n",
"4 Action,Adventure,Massively Multiplayer \n",
" genre \\\n",
"0 Action \n",
"1 Action,Adventure,Massively Multiplayer \n",
"2 Action,Adventure,Strategy \n",
"3 Action,Adventure,Massively Multiplayer \n",
"4 Action,Free to Play,Massively Multiplayer,RPG,... \n",
"\n",
" 14 \\\n",
"0 game_description \n",
"1 About This Game Developed by id software, the... \n",
"2 About This Game PLAYERUNKNOWN'S BATTLEGROUND... \n",
"3 About This Game From original BATTLETECH/Mec... \n",
"4 About This Game The post-soviet country of Ch... \n",
" game_description \\\n",
"0 About This Game Developed by id software, the... \n",
"1 About This Game PLAYERUNKNOWN'S BATTLEGROUND... \n",
"2 About This Game From original BATTLETECH/Mec... \n",
"3 About This Game The post-soviet country of Ch... \n",
"4 About This Game \n",
"\n",
" 15 \\\n",
"0 mature_content \n",
"1 NaN \n",
"2 Mature Content Description The developers de... \n",
" mature_content \\\n",
"0 NaN \n",
"1 Mature Content Description The developers de... \n",
"2 NaN \n",
"3 NaN \n",
"4 NaN \n",
"\n",
" 16 \\\n",
"0 minimum_requirements \n",
"1 Minimum:,OS:,Windows 7/8.1/10 (64-bit versions... \n",
" minimum_requirements \\\n",
"0 Minimum:,OS:,Windows 7/8.1/10 (64-bit versions... \n",
"1 Minimum:,Requires a 64-bit processor and opera... \n",
"2 Minimum:,Requires a 64-bit processor and opera... \n",
"3 Minimum:,Requires a 64-bit processor and opera... \n",
"4 Minimum:,OS:,Windows 7/8.1 64-bit,Processor:,I... \n",
"3 Minimum:,OS:,Windows 7/8.1 64-bit,Processor:,I... \n",
"4 Minimum:,OS:,Windows 7,Processor:,Intel Dual C... \n",
"\n",
" 17 18 \\\n",
"0 recommended_requirements original_price \n",
"1 Recommended:,OS:,Windows 7/8.1/10 (64-bit vers... $19.99 \n",
"2 Recommended:,Requires a 64-bit processor and o... $29.99 \n",
"3 Recommended:,Requires a 64-bit processor and o... $39.99 \n",
"4 Recommended:,OS:,Windows 10 64-bit,Processor:,... $44.99 \n",
" recommended_requirements original_price \\\n",
"0 Recommended:,OS:,Windows 7/8.1/10 (64-bit vers... $19.99 \n",
"1 Recommended:,Requires a 64-bit processor and o... $29.99 \n",
"2 Recommended:,Requires a 64-bit processor and o... $39.99 \n",
"3 Recommended:,OS:,Windows 10 64-bit,Processor:,... $44.99 \n",
"4 Recommended:,OS:,Windows 10,Processor:,Intel i... Free \n",
"\n",
" 19 \n",
"0 discount_price \n",
"1 $14.99 \n",
"2 NaN \n",
"3 NaN \n",
"4 NaN \n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"C:\\Users\\FlorianSpeicher\\AppData\\Local\\Temp\\ipykernel_38708\\931669033.py:4: DtypeWarning: Columns (12) have mixed types. Specify dtype option on import or set low_memory=False.\n",
" cars = pd.read_csv(\"./steam_games.csv\",sep=\",\",header=None)\n"
" discount_price \n",
"0 $14.99 \n",
"1 NaN \n",
"2 NaN \n",
"3 NaN \n",
"4 NaN \n"
]
}
],
"source": [
"import numpy as np\n",
"import pandas as pd\n",
"\n",
"# load data\n",
"cars = pd.read_csv(\"./steam_games.csv\",sep=\",\",header=None)\n",
"print(cars.head())\n"
"# url,types,name,desc_snippet,recent_reviews,all_reviews,release_date,developer,publisher,popular_tags,game_details,languages,achievements,genre,game_description,mature_content,minimum_requirements,recommended_requirements,original_price,discount_price\n",
"dataset = pd.read_csv(\"./steam_games.csv\",sep=\",\")\n",
"print(dataset.head())"
]
},
{
"cell_type": "markdown",
"id": "cba9750a",
"metadata": {},
"source": [
"## Preparation of the Training-Set\n",
"### Removing Uniques\n",
"We remove the following features from the Training-Set as they can uniquely identify a datapoint:\n",
"- URL\n",
"- Name of the Game\n",
"- Developer\n",
"- Publisher"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "06dedcdf",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" types desc_snippet \\\n",
"0 app Now includes all three premium DLC packs (Unto... \n",
"1 app PLAYERUNKNOWN'S BATTLEGROUNDS is a battle roya... \n",
"2 app Take command of your own mercenary outfit of '... \n",
"3 app The post-soviet country of Chernarus is struck... \n",
"4 app EVE Online is a community-driven spaceship MMO... \n",
"\n",
" recent_reviews \\\n",
"0 Very Positive,(554),- 89% of the 554 user revi... \n",
"1 Mixed,(6,214),- 49% of the 6,214 user reviews ... \n",
"2 Mixed,(166),- 54% of the 166 user reviews in t... \n",
"3 Mixed,(932),- 57% of the 932 user reviews in t... \n",
"4 Mixed,(287),- 54% of the 287 user reviews in t... \n",
"\n",
" all_reviews release_date \\\n",
"0 Very Positive,(42,550),- 92% of the 42,550 use... May 12, 2016 \n",
"1 Mixed,(836,608),- 49% of the 836,608 user revi... Dec 21, 2017 \n",
"2 Mostly Positive,(7,030),- 71% of the 7,030 use... Apr 24, 2018 \n",
"3 Mixed,(167,115),- 61% of the 167,115 user revi... Dec 13, 2018 \n",
"4 Mostly Positive,(11,481),- 74% of the 11,481 u... May 6, 2003 \n",
"\n",
" popular_tags \\\n",
"0 FPS,Gore,Action,Demons,Shooter,First-Person,Gr... \n",
"1 Survival,Shooter,Multiplayer,Battle Royale,PvP... \n",
"2 Mechs,Strategy,Turn-Based,Turn-Based Tactics,S... \n",
"3 Survival,Zombies,Open World,Multiplayer,PvP,Ma... \n",
"4 Space,Massively Multiplayer,Sci-fi,Sandbox,MMO... \n",
"\n",
" game_details \\\n",
"0 Single-player,Multi-player,Co-op,Steam Achieve... \n",
"1 Multi-player,Online Multi-Player,Stats \n",
"2 Single-player,Multi-player,Online Multi-Player... \n",
"3 Multi-player,Online Multi-Player,Steam Worksho... \n",
"4 Multi-player,Online Multi-Player,MMO,Co-op,Onl... \n",
"\n",
" languages achievements \\\n",
"0 English,French,Italian,German,Spanish - Spain,... 54.0 \n",
"1 English,Korean,Simplified Chinese,French,Germa... 37.0 \n",
"2 English,French,German,Russian 128.0 \n",
"3 English,French,Italian,German,Spanish - Spain,... NaN \n",
"4 English,German,Russian,French NaN \n",
"\n",
" genre \\\n",
"0 Action \n",
"1 Action,Adventure,Massively Multiplayer \n",
"2 Action,Adventure,Strategy \n",
"3 Action,Adventure,Massively Multiplayer \n",
"4 Action,Free to Play,Massively Multiplayer,RPG,... \n",
"\n",
" game_description \\\n",
"0 About This Game Developed by id software, the... \n",
"1 About This Game PLAYERUNKNOWN'S BATTLEGROUND... \n",
"2 About This Game From original BATTLETECH/Mec... \n",
"3 About This Game The post-soviet country of Ch... \n",
"4 About This Game \n",
"\n",
" mature_content \\\n",
"0 NaN \n",
"1 Mature Content Description The developers de... \n",
"2 NaN \n",
"3 NaN \n",
"4 NaN \n",
"\n",
" minimum_requirements \\\n",
"0 Minimum:,OS:,Windows 7/8.1/10 (64-bit versions... \n",
"1 Minimum:,Requires a 64-bit processor and opera... \n",
"2 Minimum:,Requires a 64-bit processor and opera... \n",
"3 Minimum:,OS:,Windows 7/8.1 64-bit,Processor:,I... \n",
"4 Minimum:,OS:,Windows 7,Processor:,Intel Dual C... \n",
"\n",
" recommended_requirements original_price \\\n",
"0 Recommended:,OS:,Windows 7/8.1/10 (64-bit vers... $19.99 \n",
"1 Recommended:,Requires a 64-bit processor and o... $29.99 \n",
"2 Recommended:,Requires a 64-bit processor and o... $39.99 \n",
"3 Recommended:,OS:,Windows 10 64-bit,Processor:,... $44.99 \n",
"4 Recommended:,OS:,Windows 10,Processor:,Intel i... Free \n",
"\n",
" discount_price \n",
"0 $14.99 \n",
"1 NaN \n",
"2 NaN \n",
"3 NaN \n",
"4 NaN \n"
]
}
],
"source": [
"# types,desc_snippet,recent_reviews,all_reviews,release_date,popular_tags,game_details,languages,achievements,genre,game_description,mature_content,minimum_requirements,recommended_requirements,original_price,discount_price\n",
"dataset.drop(['url', 'name', 'developer', 'publisher'], axis=1, inplace=True)\n",
"print(dataset.head())"
]
},
{
"cell_type": "markdown",
"id": "f5436c87",
"metadata": {},
"source": [
"### Structurize Text\n",
"**TODO: check if makes sense**\n",
"The dataset holds a lot of unstructured data, we use Term Frequency-Inverse Document Frequency to structurize most Text-Features.\n",
"It is important to use an new Instance for each feature so they don't overlap with each other. \n",
"\n",
"### Standardize Numbers\n",
"We standardize the prices so they can "
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "4e8b407c",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[[1. 1. 1.]]\n"
]
}
],
"source": [
"from sklearn.compose import make_column_transformer\n",
"from sklearn.feature_extraction.text import TfidfVectorizer\n",
"# types,desc_snippet,recent_reviews,all_reviews,release_date,popular_tags,game_details,languages,achievements,genre,game_description,mature_content,minimum_requirements,recommended_requirements,original_price,discount_price\n",
"column_transformer = make_column_transformer(\n",
" (TfidfVectorizer(stop_words='english'), ['desc_snippet']),\n",
" (TfidfVectorizer(stop_words='english'), ['mature_content']),\n",
" (TfidfVectorizer(stop_words='english'), ['game_description']),\n",
" (StandardScaler(), ['original_price','discount_price']) # use the same scaling for both\n",
" ('passthrough', ['price']),\n",
" #TODO: add transformer for every feature @flo @max\n",
" #TODO: check why not working\n",
")\n",
"#\n",
"dataset2 = column_transformer.fit_transform(dataset)\n",
"print(dataset2)"
]
},
{
"cell_type": "markdown",
"id": "ad84e777",
"metadata": {},
"source": [
"\n",
"### Removing Bundles\n",
"**(TODO: decide whether yes or no), not as important as i thought**\n",
"As bundles don't have clear genre(s) defined (e.g. publisher bundles )"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "base",
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
@@ -165,7 +330,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.13.5"
"version": "3.13.3"
}
},
"nbformat": 4,